Article

Uniting the Tribes: Using Textfor Marketing Insight

Jonah Berger, Ashlee Humphreys, Stephan Ludwig, Wendy W. Moe,Oded Netzer, and David A. Schweidel

AbstractWords are part of almost every marketplace interaction. Online reviews, customer service calls, press releases, marketing com-munications, and other interactions create a wealth of textual data. But how can marketers best use such data? This article provides anoverview of automated textual analysis and details how it can be used to generate marketing insights. The authors discuss how textreflects qualities of the text producer (and the context in which the text was produced) and impacts the audience or text recipient.Next, theydiscuss how text can be a powerful tool both for prediction and for understanding (i.e., insights).Then, the authorsoverviewmethodologies and metrics used in text analysis, providing a set of guidelines and procedures. Finally, they further highlight somecommon metrics and challenges and discuss how researchers can address issues of internal and external validity. They conclude with adiscussion of potential areas for future work. Along the way, the authors note how textual analysis can unite the tribes of marketing.While most marketing problems are interdisciplinary, the field is often fragmented. By involving skills and ideas from each of thesubareas of marketing, text analysis has the potential to help unite the field with a common set of tools and approaches.

Keywordscomputational linguistics,machine learning, marketing insight, interdisciplinary, natural language processing, text analysis, text mining

Online supplement: https://doi.org/10.1177/0022242919873106

The digitization of information has made a wealth of textual

data readily available. Consumers write online reviews, answer

open-ended survey questions, and call customer service repre-

sentatives (the content of which can be transcribed). Firms

write ads, email frequently, publish annual reports, and issue

press releases. Newspapers contain articles, movies have

scripts, and songs have lyrics. By some estimates, 80%–95%of all business data is unstructured, and most of that unstruc-

tured data is text (Gandomi and Haider 2015).

Such data has the potential to shed light on consumer, firm,

and market behavior, as well as society more generally. But, by

itself, all this data is just that—data. For data to be useful,

researchers must be able to extract underlying insight—to mea-

sure, track, understand, and interpret the causes and conse-

quences of marketplace behavior.

This is where the value of automated textual analysis comes

in. Automated textual analysis1 is a computer-assisted

methodology that allows researchers to rid themselves of mea-

surement straitjackets, such as scales and scripted questions,

and to quantify the information contained in textual data as it

naturally occurs. Given these benefits, the question is no longer

whether to use automated text analysis but how these tools can

best be used to answer a range of interesting questions.

This article provides an overview of the use of automated

text analysis for marketing insight. Methodologically, text

analysis approaches can describe “what” is being said and

“how” it is said, using both qualitative and quantitative inqui-

ries with various degrees of human involvement. These

Jonah Berger is Associate Professor of Marketing, Wharton School, University

of Pennsylvania, USA (email: jberger@wharton.upenn.edu). Ashlee Humphreys

is Associate Professor, Medill School of Journalism, Media, and Integrated

Marketing Communications, Northwestern University, USA (email:

a-humphreys@northwestern.edu). Stephan Ludwig is Associate Professor of

Marketing, University of Melbourne, Australia (email: stephan.ludwig@

unimelb.edu.au). Wendy W. Moe is Associate Dean of Master’s Programs,

Dean’s Professor of Marketing, and Co-Director of the Smith Analytics

Consortium, University of Maryland, USA (email: wmoe@rhsmith.umd.edu).

Oded Netzer is Professor of Business, Columbia Business School, Columbia

University, USA (email: onetzer@gsb.columbia.edu). David A. Schweidel is

Professor of Marketing, Goizueta Business School, Emory University, USA

(email: dschweidel@emory.edu).

1 Computer-aided approaches to text analysis in marketing research are

generally interchangeably referred to as computer-aided text analysis

(Pollach 2012), text mining (Netzer et al. 2012), automated text analysis

(Humphreys and Wang 2017), or computer-aided content analysis (Dowling

and Kabanoff 1996).

Journal of Marketing1-25

ª American Marketing Association 2019Article reuse guidelines:

sagepub.com/journals-permissionsDOI: 10.1177/0022242919873106

journals.sagepub.com/home/jmx

approaches consider individual words and expressions, their

linguistic relationships within a document (within-text interde-

pendencies) and across documents (across-text interdependen-

cies), and the more general topics discussed in the text.

Techniques range from computerized word counting and

applying dictionaries to supervised or automated machine

learning that helps deduce psychometric and substantive prop-

erties of text.

Within this emerging domain, we aim to make four main

contributions. First, we illustrate how text data can be used for

both prediction and understanding, to gain insight into who

produced that text, as well as how that text may impact the

people and organizations that consume it. Second, we provide a

how-to guide for those new to text analysis, detailing the main

tools, pitfalls, and challenges that researchers may encounter.

Third, we offer a set of expansive research propositions per-

taining to using text as a means to understand meaning making

in markets with a focus on how customers, firms, and societies

construe or comprehend marketplace interactions, relation-

ships, and themselves. Whereas previous treatments of text

analysis have looked specifically at consumer text (Humphreys

and Wang 2017), social media communication (Kern et al.

2016), or psychological processes (Tausczik and Pennebaker

2010), we aim to provide a framework for incorporating text

into marketing research at the individual, firm, market, and

societal levels. By necessity, our approach includes a wide-

ranging set of textual data sources (e.g., user-generated content,

annual reports, cultural artifacts, government text).

Fourth, and most importantly, we discuss how text analysis can

help “unite the tribes.” As a field, part of marketing’s value is its

interdisciplinary nature. Unlike core disciplines such as psychol-

ogy, sociology, or economics, the marketing discipline is a big

tent that allows researchers from different traditions and research

philosophies (e.g., quantitative modeling, consumer behavior,

strategy, consumer culture theory) to come together to study

related questions (Moorman et al. 2019a, b). In reality, however,

the field often seems fragmented. Rather than different rowers all

simultaneously pulling together, it often feels more like separate

tribes, each independently going off in separate directions.

Although everyone is theoretically working toward similar goals,

there tends to be more communication within groups than

between them. Different groups often speak different

“languages” (e.g., psychology, sociology, anthropology, statis-

tics, economics, organizational behavior) and use different tools,

making it increasingly difficult to have a common conversation.

However, text analysis can unite the tribes. Not only does it

involve skills and ideas from each of these areas, doing it well

requires such integration because it borrows ideas, concepts,

approaches, and methods from each tribe and incorporates them

to achieve insight. In so doing, the approach also adds value to

each of the tribes in ways that might not otherwise be possible.

We start by discussing two distinctions that are useful when

thinking about how text can be used: (1) whether text reflects or

impacts (i.e., says something about the producer or has a down-

stream impact on something else) and (2) whether text is used

for prediction or understanding (i.e., predicting something or

understanding what caused something). Next, we explain how

text may be used to unite the tribes of marketing. Then we

provide an overview of text analysis tools and methodology

and discuss key questions and measures of validity. Finally,

we close with a future research agenda.

The Universe of Text

Communication is an integral part of marketing. Not only do

firms communicate with customers, but customers communi-

cate with firms and one another. Moreover, firms communicate

with investors and society communicates ideas and values to

the public (through newspapers and movies). These communi-

cations generate text or can be transcribed into text.

A simple way to organize the world of textual data is to

think about producers and receivers—the person or organi-

zation that creates the text and the person or organization

who consumes the text (Table 1). While there are certainly

other parties that could be listed, some of the main

producers and receivers are consumers, firms, investors, and

society at large. Consumers write online reviews that are

read by other consumers, firms create annual reports that

are read by investors, and cultural producers represent soci-

etal meanings through the creation of books, movies, and

other digital or physical artifacts that are consumed by indi-

viduals or organizations.

Consistent with this distinction between text producer and

text receiver, researchers may choose to study how text reflects

or impacts. Specifically, text reflects information about, and

thus can be used to gain insight into, the text producer or one

can study how text impacts the text receiver.

Text as a Reflection of the Producer

Text reflects and indicates something about the text producer

(i.e., the person, organization, or context that created it). Cus-

tomers, firms, and organizations use language to express them-

selves or achieve desired goals, and as a result, text signals

information about the actors, organization, or society that cre-

ated it and the contexts in which it was created. Like an anthro-

pologist piecing together pottery shards to learn about a distant

civilization, text provides a window into its producers.

Take, for example, a social media post in which someone

talks about what they did that weekend. The text that person

produces provides insight into several facets. First, it provides

insight into the individual themselves. Are they introverted or

extraverted? Neurotic or conscientious? It sheds light on who

they are in general (i.e., stable traits or customer segments;

Moon and Kamakura 2017) as well as how they may be feeling

or what they may be thinking at the moment (i.e., states). In a

sense, language can be viewed as a fingerprint or signature

(Pennebaker 2011). Just like brush strokes or painting style can

be used to determine who painted a particular painting,

researchers use words and linguistic style to infer whether a

play was written by Shakespeare, or if a person is depressed

(Rude, Gortner, and Pennebaker 2004) or being deceitful

2 Journal of Marketing XX(X)

(Ludwig et al. 2016). The same is true for groups, organiza-

tions, or institutions. Language reflects something about who

they are and thus provides insight into what they might do in

the future.

Second, text can provide insight into a person’s attitudes

toward or relationships with other attitude objects—whether

that person liked a movie or hated a hotel stay, for example,

or whether they are friends or enemies with someone. Lan-

guage used in loan applications provides insight into whether

people will default (Netzer, Lemaire, and Herzenstein 2019),

language used in reviews can provide insight into whether they

are fake (Anderson and Simester 2014; Hancock et al. 2007;

Table 1. Text Producers and Receivers.

TextProducers

Text Receivers

Consumers Firms Investors Institutions/Society

Consumers � Online reviews (Anderson andSimester 2014; Chen and Lurie 2013;Fazio and Rockledge 2015a; Kronrodand Danziger 2013a; Lee and Bradlow2011; Liu, Lee, and Srinivasan 2019a;Melumad, Inman, and Pham 2019;Moon and Kamakura 2017; Puranam,Narayan, and Kadiyali 2017)� Social media (Hamilton, Schlosser, and

Chen 2017a; Netzer et al. 2012;Villarroel Ordenes et al. 2017)� Offline word of mouth (Berger and

Schwartz 2011a; Mehl and Pennebaker2003a)

� Forms and applications(Netzer, Lemaire, andHerzenstein 2019)� Idea-generation contexts

(Bayus 2013a; Toubia andNetzer 2017)� Social media/brand

communities (Herhausenet al. 2019)� Consumer complaints

(Ma, Baohung, and Kekre2015)� Customer language on

service calls� Tweeting at companies

(Liu, Singh, and Srinivasan2016a)

� Stock market reactions toconsumer text (Bollen,Mao, and Zeng 2011;Tirunillai and Tellis 2012)� Protests� Petitions

� Crowdsourcingknowledge� Letters to the editor� Online comments

section� Activism (e.g.,

organizing politicalmovements andmarches)

Firms � Owned media (e.g., company websiteand social media; Villarroel Ordeneset al. 2018)� Advertisements (Fossen and

Schweidel 2017a, 2019; Liaukonyte,Teixeira, and Wilbur 2015a; Rosa et al.1999; Stewart and Furse 1986)� Customer service agents (Packard and

Berger 2019; Packard, Moore, andMcFerran 2018)� Packaging, including labels� Text used in instructions

� Trade publications(Weber, Heinze, andDeSoucey 2008a)� Interfirm communication

emails (Ludwig et al.2016)� White papers

� Financial reports(Loughran and McDonald2016)� Corporate

communications (Hobson,Mayhew, andVenkatachalam 2012)� Chief executive officer

letters to shareholders(Yadav, Prabhu, andChandy 2007

� Editorials by firmstakeholders� Interviews with

business leaders

Investors � Letters to shareholders(Yadav, Prabhu, andChandy 2007)� Shareholder feedback

(Wies et al. 2019)

� Sector reports

Institutions/society

� News content (Berger, Kim, andMeyer 2019; Berger and Milkman2012; Humphreys 2010)� Movies (Berger, Moe, and Schweidel

2019; Eliashberg, Hui, and Zhang 2007,2014; Toubia et al. 2019)� Songs (Berger and Packard 2018;

Packard and Berger 2019)� Books (Akpinar and Berger 2015;

Sorescu et al. 2018a)

� Business section� Specialty magazines (e.g.,

Wired, Harvard BusinessReview)

� Wall Street Journal� Fortune� Various forms of

investment advice thatcome from media

� Governmentdocuments, hearings,and memoranda(Chappell et al. 1997a)� Forms of public

dialogue or debate

aReference appears in the Web Appendix.

Berger et al. 3

Ott, Cardie, and Hancock 2012), and language used by political

candidates could be used to study how they might govern in the

future.

These same approaches can also be used to understand

leaders, organizations, or cultural elites through the text they

produce. For example, the words a leader uses reflect who

they are as an individual, their leadership style, and their

attitudes toward various stakeholders. The language used in

ads, on websites, or by customer service agents reflects infor-

mation about the company those pieces of text represent.

Aspects such as brand personality (Opoku, Abratt, and Pitt

2006), how much a firm is thinking about its customers (Pack-

ard and Berger 2019), or managers’ orientation toward end

users (Molner, Prabhu, and Yadav 2019) can be understood

through text. Annual reports provide insight into how well a

firm is likely to perform in the future (Loughran and McDo-

nald 2016).

Yet beyond single individuals or organizations, text can also

be aggregated across creators to study larger social groups or

institutions. Given that texts reflect information about the peo-

ple or organizations that created them, grouping people or

organizations together on the basis of shared characteristics

can provide insight into the nature of such groups and differ-

ences between them. Analyzing blog posts, for example, can

shed light on how older and younger people view happiness

differently (e.g., as excitement vs. peacefulness; Mogilner,

Kamvar, and Aaker 2011). In a comparison of newspaper arti-

cles and press releases about different business sectors, text can

be used to understand the creation and spread of globalization

discourse (Fiss and Hirsch 2005). Customers’ language use

further gives insight into the consumer sentiment in online

brand communities (Homburg, Ehm, and Artz 2015).

More broadly, because texts are shaped by the contexts (e.g.,

devices, cultures, time periods) in which they were produced,

they also reflect information about these contexts. In the case of

culture, U.S. culture values high-arousal positive affective

states more than East Asian cultures (Tsai 2007), and these

differences may show up in the language these different groups

use. Similarly, whereas members of individualist cultures tend

to use first-person pronouns (e.g., “I”), members of collectivist

cultures tend to use a greater proportion of third-person pro-

nouns (e.g., “we”).

Across time, researchers were able to examine whether the

national mood changed after the September 11 attacks by

studying linguistic markers of psychological change in online

diaries (Cohn, Mehl, and Pennebaker 2004). The language

used in news articles, songs, and public discourse reflects

societal attitudes and norms, and thus analyzing changes over

time can provide insight into aspects such as attitudes toward

women and minorities (Boghrati and Berger 2019; Garg et al.

2018) or certain industries (Humphreys 2010). Journal articles

provide a window into the evolution of topics within acade-

mia (Hill and Carley 1999). Books and movies serve as sim-

ilar cultural barometers and could be used to shed light on

everything from cultural differences in customs to changes in

values over time.

Consequently, text analysis can provide insights that may

not be easily (or cost-effectively) obtainable through other

methods. Companies and organizations can use social listening

(e.g., online reviews and blog posts) to understand whether

consumers like a new product, how customers feel about their

brand, what attributes are relevant for decision making, or what

other brands fall in the same consideration set (Lee and Bra-

dlow 2011; Netzer et al. 2012). Regulatory agencies can deter-

mine adverse reactions to pharmaceutical drugs (Feldman et al.

2015; Netzer et al. 2012), public health officials can gauge how

bad the flu will be this year and where it will hit the hardest

(Alessa and Faezipour 2018), and investors can try to predict

the performance of the stock market (Bollen, Mao, and Zeng

2011; Tirunillai and Tellis 2012).

Text’s Impact on Receivers

In addition to reflecting information about the people, organi-

zations, or society that created it, text also impacts or shapes the

attitudes, behavior, and choices of the audience that consumes

it. For example, take the language used by a customer service

agent. While that language certainly reflects something about

that agent (e.g., their personality, how they are feeling that

day), how they feel toward the customer, and what type of

brand they represent, that language also impacts the customer

who receives it (Packard and Berger 2019; Packard, Moore,

and McFerran 2018). It can change customer attitudes toward

the brand, influence future purchase, or affect whether custom-

ers talk about the interaction with their friends. In that sense,

language has a meaningful and measurable impact on the

world. It has consequences.

This can be seen in a myriad of different contexts. Ad copy

shapes customers’ purchase behavior (Stewart and Furse

1986), newspaper language changes customers’ attitudes

(Humphreys and LaTour 2013), trade publications and con-

sumer magazines shift product category perceptions (e.g.,

Rosa et al. 1999), movie scripts shape audience reactions

(Berger, Kim, and Meyer 2019; Eliashberg, Hui, and Zhang

2014; Reagan et al. 2016), and song lyrics shape song market

success (Berger and Packard 2018; Packard and Berger 2019).

The language used in political debates shapes which topics get

attention (Berman et al. 2019), the language used in conversa-

tion shapes interpersonal attitudes (Huang et al. 2017), and the

language used in news articles shapes whether people read

(Berger, Moe, and Schweidel, 2019b) or share (Berger and

Milkman 2012) them.

Firms’ language choice has impact as well. For example,

nuances in language choices by firms when responding to cus-

tomer criticism online directly impacts consumers and, thus,

the firms’ success in containing social media firestorms (Her-

hausen et al. 2019). Language used in YouTube ads is corre-

lated with their virality (Tellis et al. 2019). Shareholder

complaints about nonfinancial concerns and topics that receive

high media attention substantially increase firms’ advertising

investments (Wies et al. 2019).

4 Journal of Marketing XX(X)

Note that while the distinction between text reflecting and

impacting is a useful one, it is not an either/or. Text almost

always simultaneously reflects and impacts. Text always

reflects information about the actor or actors that created it,

and as long as some audience consumes that text, it also

impacts that audience.

Despite this relationship, researchers studying reflection

versus impact tend to use text differently. Research that exam-

ines what the text reflects often treats it as a dependent variable

and investigates how it relates to the text creator’s personality,

the social groups they belong to, or the time period or culture in

which it was created.

Research that examines how text impacts others often treats

it as an independent variable, examining if and how text shapes

outcomes such as purchase, sharing, or engagement. In this

framework, textual elements are linked with outcomes that are

believed to be theoretical consequences of the textual compo-

nents or some latent variable that they are thought to represent.

Contextual Influences on Text

Importantly, text is also shaped by contextual factors; thus, to

better understand its meaning and impact, it is important to

understand the broader situation in which it was produced.

Context can affect content in three ways: through technical

constraints and social norms of the genre, through shared

knowledge specific to the speaker and receiver, and through

prior history.

First, different types of texts are influenced by formal and

informal rules and norms that shape the content and expecta-

tions about the message. For example, newspaper genres such

as opinion pieces or feature stories will contain a less

“objective” point of view than traditional reporting (Ljung

2000). Hotel comment cards and other feedback are usually

dominated by more extreme opinions. On Snapchat and other

social media platforms, messages are relatively recent, short,

and often ephemeral. In contrast, online reviews can be longer

and are often archived dating back several years. Synchronic

text exchanges, in which two individuals interactively commu-

nicate in real time may be more informal and contain dialogue

of short statements and phatic responses (i.e., communication

such as “Hi,” which serves a social function) that indicate

affiliation rather than semantic content (Kulkarni 2014). Some

genres (e.g., social media) are explicitly public, whereas on

others, such as blogs, information that is more private may

be conveyed.

Text is also shaped by technological constraints (e.g., the

ability to like or share) and physical constraints (e.g., charac-

ter length limitations). Tweets, for example, necessarily have

280 characters or fewer, which may shape the ways in which

they are used to communicate. Mobile phones have con-

straints on typing and may shape the text that people produce

on them (Melumad, Inman, and Pham 2019; Ransbotham,

Lurie, and Liu 2019).

Second, the relationship between the text producer and con-

sumer may affect what is said (or, more often, unsaid). If the

producer and consumer know each other well, text may be

relatively informal (Goffman 1959) and lack explicit informa-

tion that a third party would need to make sense of the con-

versation (e.g., past events, known likes/dislikes). If both have

an understanding of the goal of the communication (e.g., that

the speaker wants to persuade the receiver), this may shape the

content but be less explicit.

These factors are important to understand when interpret-

ing the content of the text itself. Content has been shown to

be shaped by the creator’s intended audience (Vosoughi,

Roy, and Aral 2018) and anticipated effects on the receiver

(Barasch and Berger 2014). Similarly, what consumers share

with their best friend may be different (e.g., less impacted

by self-presentational motivations) than what they post

online for everyone to see.2 Firms’ annual reports may be

shaped by the goals of appearing favorably to the market.

What people say on a customer service call may be driven

by the goal of getting monetary compensation. Consumer

protests online are meant to inspire change, not merely

inform others.

Finally, history may affect the content of the text. In mes-

sage boards, prior posts may shape future posts; if someone

raised a point in a previous post, the respondent will most likely

refer to the point in future posts. If retweets are included in an

analysis, this will bias content toward most circulated posts.

More broadly, media frames such as #metoo or #blacklives-

matter might make some concepts or facts more accessible to

speakers and therefore more likely to emerge in text, even if

seemingly unrelated (McCombs and Shaw 1972; Xiong, Cho,

and Boatwright 2019).

Using Text for Prediction VersusUnderstanding

Beyond reflecting information about the text creator and

shaping outcomes for the text recipient, another useful dis-

tinction is whether text is used for prediction or

understanding.

Prediction

Some text research is predominantly interested in prediction.

Which customer is most likely to default on their loan (Netzer,

Lemaire, and Herzenstein 2019)? Which movie will sell the

most tickets (Eliashberg et al. 2014)? How will the stock mar-

ket perform (Bollen, Mao, and Zeng 2011; Tirunillai and Tellis

2012)? Whether focusing on individual-, firm-, or market-level

outcomes, the goal is to predict with the highest degree of

accuracy. Such work often takes many textual features and uses

machine learning or other methods to combine these features in

a way that achieves the best prediction. The authors care less

2 Note that intermediaries can amplify (e.g., retweet) an original message and

may have different motivations than the text producer.

Berger et al. 5

about any individual feature and more about how the set of

observable features can be combined to predict an outcome.

The main difficulty involved with using text for predictions

is that text can generate hundreds and often thousands of fea-

tures (words) that are all potential predictors for the outcome of

interest. In some cases, the number of predictors is larger than

the number of observations, making traditional statistical pre-

dictive models largely impractical. To address this issue,

researchers often resort to machine learning–type methods, but

overfitting needs to be carefully considered. In addition, infer-

ence with respect to the role of each word in the prediction can

be difficult. Methods such as feature importance weighing can

help extract some inference from these predictive models.

Understanding

Other research is predominantly interested in using text for

understanding. How does the language consumers use shape

word of mouth’s impact (Packard and Berger 2017)? Why do

some online posts get shared, songs become popular, or

brands engender greater loyalty? How do cultural attitudes

or business practices change? Whether focusing on

individual-, firm-, or market-level outcomes, the goal is to

understand why or how something occurred. Such work often

involves examining only one or a small number of textual

features or aspects that link to underlying psychological or

sociological processes and aims to understand which features

are driving outcomes and why.

One challenge with using textual data for understanding

is drawing causal inferences from observational data. Con-

sequently, work in this area may augment field data with

experiments to allow key independent variables to be

manipulated. Another challenge is interpreting relationships

with textual features (we discuss this further in the closing

section). Songs that use more second-person pronouns are

more popular (Packard and Berger 2019), for example, but

this relationship alone does not necessarily explain why this

is the case; second-person pronouns may indicate several

things. Consequently, deeper theorizing, examination of

links observed in prior research, or further empirical work

is often needed.

Note that research that can use either a prediction or

understanding lens to study either what text reflects or what

it impacts. On the prediction side, researchers interested in

what text reflects could use it to predict states or traits of

the text creator such as customer satisfaction, likelihood of

churn, or brand personality. Researchers interested in the

impact of text could predict how text will shape outcomes

such as reading behavior, sharing, or purchase among con-

sumers of that text.

On the understanding side, someone interested in what text

reflects could use it to shed light on why people might use

certain personal pronouns when they are depressed or why

customers might use certain types of emotional language

when they are talking to customer service. Someone inter-

ested in the impact of text could use it to understand why text

that evokes different emotions might be more likely to be read

or shared.

Furthermore, while most research tends to focus on either

prediction or understanding, some work integrates both

aspects. Netzer, Lemaire, and Herzenstein (2019), for example,

both use a range of available textual features to predict whether

a given person will default on a loan and analyze the specific

language used by people who tend to default (e.g., language

used by liars).

Uniting the Tribes of Marketing

Regardless of whether the focus is on text reflection versus

impact, or prediction versus understanding, doing text analysis

well requires integrating skills, techniques, and substantive

knowledge from different areas of marketing. Furthermore,

textual analysis opens up a wealth of opportunity for each of

these areas as well.

Take consumer behavior. While hypothetical scenarios can

be useful, behavioral economics has recently gotten credit for

many applications of social or cognitive psychology because

these researchers have demonstrated phenomena in the field.

Given concerns about replication, researchers have started to

look for new tools that enable them to ensure validity and

increase relevance to external audiences. Previously, use of

secondary data was often limited because it addressed the

“what” but not the “why” (i.e., what people bought or did, but

not why they did so). But text can provide a window into the

underlying process. Online reviews, for example, can be used

to understand why someone bought one thing rather than

another. Blog posts can help marketers understand consider-

ation sets (Lee and Bradlow 2011; Netzer et al. 2012) and the

customer journey (Li and Du 2011). Text even helps address

the age-old issue of telling more than we can know (Nisbett and

Wilson 1977). While people may not always know why they

did something, their language often provides traces of explana-

tion (Pennebaker 2011), even beyond what they can con-

sciously articulate.

This richness is attractive to more than just behavioral

researchers. Text opens a large-scale window into the world

of “why” in the field and does so in a scalable manner. Quan-

titative modelers are always looking for new data sources and

tools to explain and predict behavior. Unstructured data pro-

vides a rich set of predictors that are often readily available, at

large scale, and able to be combined with structured measures

as either dependent variables or independent variables. Text,

through product reviews, user-driven social media activity, and

firm-driven marketing efforts, provides data in real time that

can shed light on consumer needs/preferences. This offers an

alternative or supplement to traditional marketing research

tools. In many cases, text can be retraced to an individual,

allowing distinction between individual differences and

dynamics. It also offers a playground where new methodolo-

gies from other disciplines can be applied (e.g., deep learning;

LeCun, Bengio, and Hinton 2015; Liu et al. 2019).

6 Journal of Marketing XX(X)

Marketing strategy researchers want logic by which business

can achieve its marketing objectives and to better understand

what affects organizational success. A primary challenge to

these researchers is to obtain reliable and generalizable survey

or field data about factors that lie deep in the firm’s culture and

structure or that are housed in the mental models and beliefs of

marketing leaders and employees. Text analysis offers an objec-

tive and systematic solution to assess constructs in naturally

occurring data (e.g., letters to shareholders, press releases, patent

text, marketing messages, conference calls with analysts) that

may be more valid. Likewise, marketing strategy scholars often

struggle with valid measures of a firm’s marketing assets, and

text may be a useful tool to understand the nature of customer,

partner, and employee relationships and the strength of brand

sentiments. For example, Kubler, Colicev, and Pauwels (2017)

use dictionaries and support vector machine methods to extract

sentiment and relate it to consumer mindset metrics.

Scholars who draw from anthropology and sociology have

long examined text through qualitative interpretation and con-

tent analysis. Consumer culture theory–oriented marketing

researchers are primarily interested in understanding underly-

ing meanings, norms, and values of consumers, firms, and

markets in the marketplace. Text analysis provides a tool for

quantifying qualitative information to measure changes over

time or make comparisons between groups. Sociological and

anthropological researchers can use automated text analysis to

identify important words, locate themes, link them to text seg-

ments, and examine common expressions in their context. For

example, to understand consumer taste practices, Arsel and

Bean (2013) use text analysis to first identify how consumers

talk about different taste objects, doings, and meanings in their

textual data set (comments on a website/blog) before analyzing

the relationship between these elements using interview data.

For marketing practitioners, textual analysis unlocks the

value of unstructured data and offers a hybrid between quali-

tative and quantitative marketing research. Like qualitative

research, it is rich, exploratory, and can answer the “why,” but

like quantitative research, it benefits from scalability, which

often permits modeling and statistical testing. Textual analysis

enables researchers to explore open-ended questions for which

they do not know the range of possible answers a priori. With

text, scholars can answer questions that they did not ask or for

which they did not know the right outcome measure. Rather

than forcing on participants a certain scale or set of outcomes

from which to select, for example, marketing researchers can

instead ask participants broad questions, such as why they like

or dislike something, and then use topic modeling tools such as

latent Dirichlet allocation (LDA; explained in detail subse-

quently) to discover the key underlying themes.

Importantly, while text analysis offers opportunities for a

variety of research traditions, such opportunities are more

likely to be realized when researchers work across traditional

subgroups. That is, the benefits of computer-aided text analysis

are best realized if we include both quantitative, positivist

analyses of content and qualitative, interpretive analyses of

discourse. Quantitative researchers, for example, have the

skills to build the right statistical models, but they can benefit

from behavioral and qualitative researchers’ ability to link

words to underlying psychological or social processes as well

as marketing strategy researchers’ understanding of organiza-

tional and marketing activities driving firm performance. This

is true across all of the groups.

Thus, to really extract insights from textual data, research

teams must have the interpretative skills to understand the

meaning of words, the behavioral skills to link them to under-

lying psychological processes, the quantitative skills to build

the right statistical models, and the strategy skills to understand

what these findings mean for firm actions and outcomes. We

outline some potential areas for fruitful collaboration in

“Future Research Agenda” section.

Text Analysis Tools, Methods, and Metrics

Given the recent work using text analysis to derive marketing

insight, some researchers may wonder where to start. This

section reviews methodologies often used in text-based

research. These include techniques needed to convert text into

constructs in the research process as well as procedures needed

to incorporate extracted textual information into subsequent

modeling and analyses. The objective of this section is not to

provide a comprehensive tutorial but, rather, to expose the

reader to available techniques, discuss when different methods

are appropriate, and highlight some of the key considerations in

applying each method.

The process of text analysis involves several steps: (1) data

preprocessing, (2) performing a text analysis of the resulting

data, (3) converting the text into quantifiable measures, and

(4) assessing the validity of the extracted text and measures.

Each of these steps may vary depending on the research objec-

tive. Table 2 provides a summary of the different steps

involved in the text analysis process from preprocessing to com-

monly used tools and measures and validation approaches. Table

2 can serve as a starter kit for those taking their first steps with

text analysis.

Data Preprocessing

Text is often unstructured and “messy,” so before any formal

analyses can take place, researchers must first preprocess the

text itself. This step provides structure and consistency so that

the text can be used systematically in the scientific process.

Common software tools for text analysis include Python (https://

www.nltk.org/) and R (https://cran.r-project.org/web/packages/

quanteda/quanteda.pdf, https://quanteda.io/). For both software

platforms, a set of relatively easy-to-use tools has been devel-

oped to perform most of the data preprocessing steps. Some

programs, such as Linguistic Inquiry and Word Count (LIWC;

Tausczik and Pennebaker 2010) and WordStat (Peladeau 2016),

require minimal preprocessing. We detail the data preprocessing

steps next (for a summary of the steps, see Table 3).

Berger et al. 7

Data acquisition. Data acquisition can be well defined if the

researcher is provided with a set of documents (e.g., emails,

quarterly reports, a data set of product reviews) or more open-

ended if the researcher is using a web scraper (e.g., Beautiful

Soup) that searches the web for instances of a particular topic

or a specific product. When scraping text from public sources,

researchers should abide by the legal guidelines for using the

data for academic or commercial purposes.

Tokenization. Tokenization is the process of breaking the text into

units (often words and sentences). When tokenizing, the

researcher needs to determine the delimiters that define a token

(space, period, semicolon, etc.). If, for example, a space or a period

is used to determine a word, it may produce some nonsensical

tokens. For example, “the U.S.” may be broken to the tokens “the,”

“U,” and “S.” Most text-mining software has smart tokenization

procedures to alleviate such common problems, but the

researcher should pay close attention to instances that are spe-

cific to the textual corpora. For cases that include paragraphs or

threads, depending on the research objective, the researcher

may wish to tokenize these larger units of text as well.

Cleaning. HTML tags and nontextual information, such as

images, are cleaned or removed from the data set. The cleaning

needs may depend on the format in which the data was provided/

extracted. Data extracted from the web often requires heavier

cleaning due to the presence of HTML tags. Depending on the

purpose of the analysis, images and other nontextual information

may be retained. Contractions such as “isn’t” and “can’t” need to

be expanded at this step. In this step, researchers should also be

mindful of and remove phrases automatically generated by com-

puters that may occur within the text (e.g., “html”).

Removing stop words. Stop words are common words such as “a”

and “the” that appear in most documents but often provide no

significant meaning. Common text-mining tools (e.g., the tm,

quanteda, tidytext, and tokenizers package in R; the Natural

Language Toolkit package in Python; exclusion words in

WordStat) have a predefined list of such stop words that can

be amended by the researcher. It is advisable to add common

words that are specific to the domain (e.g., “Amazon” in a

corpora of Amazon reviews) to this list. Depending on the

research objective, stop words can sometimes be very mean-

ingful, and researchers may wish to retain them for their anal-

ysis. For example, if the researcher is interested in extracting

not only the content of the text but also writing style (e.g.,

Packard, Moore, and McFerran 2018), stop words can be very

informative (Pennebaker 2011).

Spelling. Most text-mining packages have prepackaged spellers

that can help correct spelling mistakes (e.g., the Enchant spel-

ler). In using these spellers, the researcher should be aware of

language that is specific to the domain and may not appear in

the speller—or even worse, that the speller may incorrectly

“fix.” Moreover, for some analyses the researcher may want

to record the number of spelling mistakes as an additional

Table 2. The Text Analysis Workflow.

Data Preprocessing Common Tools Measurement Validity

� Data acquisition:Obtain or download(often in an HTMLformat) text.� Tokenization: Break

text into units (oftenwords and sentences)using delimiters (e.g.,periods).� Cleaning: Remove

nonmeaningful text (e.g.,HTML tags) andnontextual information.� Removing stop words:

Eliminate commonwords such as “a” or“the” that appear inmost documents.� Spelling: Correct

spelling mistakes usingcommon spellers.� Stemming and

lemmatization: Reducewords into theircommon stem or lemma.

� Entity extraction: Tools used toextract the meaning of one wordat a time or simple cooccurrenceof words. These tools includedictionaries; part-of-speechclassifiers; many sentiment analysistools; and, for complex entities,machine learning tools.� Topic modeling: Topic modeling

can identify the general topics(described as a combination ofwords) that are discussed in abody of text. Common toolsinclude LDA and PF.� Relation extraction: Going

beyond entity extraction, theresearcher may be interested inidentifying textual relationshipsamong extracted entities. Relationextraction often requires the useof supervised machine learningapproaches.

� Count measures: The set ofmeasures used to represent thetext as count measures. The tf-idfmeasure allows the researcher tocontrol for the popularity of theword and the length of thedocument.

� Similarity measures: Cosinesimilarity and the Jaccard indexare often used to measure thesimilarity of the text betweendocuments.

� Accuracy measures: Often usedrelative to human-coded orexternally validated documents.The measures of recall, precision,F1, and the area under the curveof the receiver operatingcharacteristic curve are oftenused.

� Readability measures: Measuressuch as the simple measure ofgobbledygook (SMOG) are usedto assess the readability level ofthe text.

� Internal Validity– Construct: Dictionary

validation and sampling-and-saturation procedures ensurethat constructs are correctlyoperationalized in text.

– Concurrent: Compareoperationalizations with priorliterature.

– Convergent: Multipleoperationalizations of keyconstructs.

– Causal: Control for factorsrelated to alternativehypotheses.

� External Validity– Predictive: Use conclusions to

predict key outcome variable(e.g., sales, stock price).

– Generalizability: Replicateeffects in other domains.

– Robustness: Test conclusionson holdout samples (k-fold);compare different categorieswithin the data set.

Note: PF ¼ Poisson factoring.

8 Journal of Marketing XX(X)

textual measure reflecting important states or traits of the com-

municator (e.g., Netzer, Lemaire, and Herzenstein 2019).

Stemming and lemmatization. Stemming is the process of reducing

the words into their word stem. Lemmatization is similar to stem-

ming, but it returns the proper lemma as opposed to the word’s root,

which may not be a meaningful word. For example, with stem-

ming, the entities “car” and “cars” are stemmed to “car,” but

“automobile” is not. In lemmatization, the words “car,” “cars,”

and “automobile” are all reduced to the lemma “automobile.”

Several prepackaged stemmers exist in most text-mining tools

(e.g., the Porter stemmer). Similar to stop words, if the goal of the

analysis is to extract the writing style, one may wish to skip the

stemming step, because stemming often masks the tense used.

Text Analysis Extraction

Once the data has been preprocessed, the researcher can start

analyzing the data. One can distinguish between the extraction

of individual words or phrases (entity extraction), the extrac-

tion of themes or topics from the collective set of words or

phrases in the text (topic extraction), and the extraction of

relationships between words or phrases (relation extraction).

Table 4 highlights these three types of analysis, the typical

research questions investigated with each approach, and some

commonly used tools.

Entity (word) extraction. At the most basic level, text mining has

been used in marketing to extract individual entities (i.e., count

words) such as person, location, brands, product attributes, emo-

tions, and adjectives. Entity extraction is probably the most

commonly used text analysis approach in marketing academia

and practice, partly due to its relative simplicity. It allows the

researcher to explore both what was written (the content of the

words) as well as how it was written (the writing style). Entity

extraction can be used (1) to monitor discussions on social media

(e.g., numerous commercial companies offer buzz monitoring

services and use entity extraction to track how frequently a brand

is being mentioned across alternative social media), (2) to gen-

erate a rich set of entities (words) to be used in a predictive

model (e.g., which words or entities are associated with fake

or fraudulent statements), and (3) as input to be used with dic-

tionaries to extract more complex forms of textual expressions,

such as a particular concept, sentiment, emotion, or writing style.

In addition to programming languages such as Python and

R’s tm tool kits, software packages such as WordStat make it

possible to extract entities without coding. Entity extraction

can also serve as input in commonly used dictionaries or lex-

icons. Dictionaries (i.e., a predefined list of words, such as a list

of brand names) are often used to classify entities into the

categories (e.g., concepts, brands, people, categories, loca-

tions). In more formal text, capitalization can be used to help

Table 3. Data Preprocessing Steps.

Data Processing Step Issues to Consider Illustration

Data acquisition � Is the data readily available in textual format or does theresearch needs to use a web scraper to find the data?� What are the legal guidelines for using the data

(particularly relevant for web-scraped data)?

Tweets mentioning different brands from the samecategory during a particular time frame aredownloaded from Twitter.

Tokenization � What is the unit of analysis (word, sentence, thread,paragraph)?� Use smart tokenization for delimiters and adjust to

specific unique delimiters found in the corpora.

The unit of analysis is the individual tweet. The wordsin the tweet are the tokens of the document.

Cleaning � Web-scraped data often requires cleaning of HTMLtags and other symbols.� Depending on the research objective, certain textual

features (e.g., advertising on the page) may or may notbe cleaned.� Expansion of contractions such as “isn’t” to “is not.”

URLs are removed and emojis/emoticons areconverted to words.

Removing stop word � Use a stop word list available by the text-miningsoftware, but adapt it to a specific application by adding/removing relevant stop words.� If the goal of the analysis is to extract writing style, it is

advisable to keep all/some of the stop words.

Common words are removed. The remaining textcontains brand names, nouns, verbs, adjectives,and adverbs.

Spelling � Can use commonly used spellers in text-miningpackages (e.g., the Enchant speller).� Language that is specific to the domain may be

erroneously coded as a spelling mistake.� May wish to record the number of spelling mistakes as

an additional textual measure.

Spelling mistakes are removed, enabling analysis intoconsumer perceptions (manifest through wordchoice) of different brands.

Stemming and lemmatization � Can use commonly used stemmers in text-miningpackages (e.g., Porter stemmer).� If the goal of the analysis is to extract writing style,

stemming can mask the tense used.

Verbs and nouns are “standardized” by reducing totheir stem or lemma.

Berger et al. 9

Tab

le4.

Tax

onom

yofT

ext

Anal

ysis

Tools

.

Appro

ach

Com

mon

Tools

Res

earc

hQ

ues

tions

Ben

efits

Lim

itat

ions

and

Com

ple

xitie

sM

arke

ting

Exam

ple

s

Entity

(word

)ex

trac

tion:

Extr

acting

and

iden

tify

ing

asi

ngl

ew

ord

/n-g

ram

�N

amed

entity

extr

action

(NER

)to

ols

(e.g

.,St

anfo

rdN

ER

)�

Dic

tionar

ies

and

lexic

ons

(e.g

.,LI

WC

,EL

2.0

,Se

ntiSt

rengt

h,

VA

DER

)�

Rule

-bas

edcl

assi

ficat

ion

�Li

ngu

istic-

bas

edN

LPto

ols

�M

achin

ele

arnin

gcl

assi

ficat

ion

tools

(conditio

nal

random

field

s,hid

den

Mar

kov

model

s,dee

ple

arnin

g)

�Bra

nd

buzz

monitori

ng

�Pre

dic

tive

model

sw

her

ete

xt

isan

input

�Extr

acting

psy

cholo

gica

lst

ates

and

trai

ts�

Sentim

ent

anal

ysis

�C

onsu

mer

and

mar

ket

tren

ds

�Pro

duct

reco

mm

endat

ions

�C

anex

trac

ta

larg

enum

ber

of

entities

�C

anunco

ver

know

nen

tities

(peo

ple

,bra

nds,

loca

tions)

�C

anbe

com

bin

edw

ith

dic

tionar

ies

toex

trac

tse

ntim

ent

or

lingu

istic

styl

es�

Rel

ativ

ely

sim

ple

touse

�C

anbe

unw

ield

ydue

toth

ela

rge

num

ber

ofen

tities

extr

acte

d�

Som

een

tities

hav

em

ultip

lem

eanin

gsth

atar

ediff

icult

toex

trac

t(e

.g.,

the

laundry

det

erge

nt

bra

nd

“All”

)�

Slan

gan

dab

bre

viat

ions

mak

een

tity

extr

action

more

diff

icult

inso

cial

med

ia�

Mac

hin

ele

arnin

gto

ols

may

requir

ela

rge

hum

an-c

oded

trai

nin

gdat

a�

Can

be

limited

for

sentim

ent

anal

ysis

�Le

ean

dBra

dlo

w(2

011)

�Ber

ger

and

Milk

man

(201

2)�

Ghose

etal

.(2

012)a

�T

irunill

aian

dT

ellis

(2012)

�H

um

phre

ysan

dT

hom

pso

n(2

014)a

�Ber

ger,

Moe,

and

Schw

eidel

(2019)

�Pac

kard

,M

oore

,an

dM

cFer

ran

(2018)

Topic

extr

action:

Extr

acting

the

topic

dis

cuss

edin

the

text

�LS

A�

LDA

�PF

�LD

A2ve

cw

ord

embed

din

g

�Su

mm

ariz

ing

the

dis

cuss

ion

�Id

entify

ing

consu

mer

and

mar

ket

tren

ds

�Id

entify

ing

cust

om

ernee

ds

�T

opic

soft

enpro

vide

use

ful

sum

mar

izat

ion

ofth

edat

a�

Dat

are

duct

ion

per

mits

the

use

oftr

aditio

nal

stat

istica

lm

ethods

insu

bse

quen

tan

alys

is�

Eas

y-to

-ass

ess

dyn

amic

s

�T

he

inte

rpre

tation

ofth

eto

pic

sca

nbe

chal

lengi

ng

�N

ocl

ear

guid

ance

on

the

sele

ctio

nofth

enum

ber

ofto

pic

s�

Can

be

diff

icult

with

short

text

(e.g

.,tw

eets

)

�T

irunill

aian

dT

ellis

(2014)

�Busc

hke

nan

dA

llenby

(2016)

�Pura

nam

,N

aray

an,an

dK

adiy

ali(2

017)

�Ber

ger

and

Pac

kard

(2018)

�Li

uan

dT

oubia

(2018)

�T

oubia

etal

.(2

019)

�Z

hong

and

Schw

eidel

(2019)

�A

nsa

ri,Li

,an

dY

ang

(2018)a

�T

imosh

enko

and

Hau

ser

(2019)

�Li

u,Si

ngh

,an

dSr

iniv

asan

(2016)a

�Li

u,Le

e,an

dSr

iniv

asan

(2019)a

Rel

atio

nex

trac

tion:

Extr

acting

and

iden

tify

ing

rela

tionsh

ips

among

word

s

�C

o-o

ccurr

ence

ofen

tities

�H

andw

ritt

enru

le�

Super

vise

dm

achin

ele

arnin

g�

Dee

ple

arnin

g�

Word

2ve

cw

ord

embed

din

g�

Stan

ford

Sente

nce

and

Gra

mm

atic

alD

epen

den

cyPar

ser

�M

arke

tm

appin

g�

Iden

tify

ing

pro

ble

ms

men

tioned

with

spec

ific

pro

duct

feat

ure

s�

Iden

tify

ing

sentim

ent

for

afo

cal

entity

�Id

entify

ing

whic

hpro

duct

attr

ibute

sar

em

entioned

posi

tive

ly/n

egat

ivel

y�

Iden

tify

ing

even

tsan

dco

nse

quen

ces

(e.g

.,cr

isis

)fr

om

consu

mer

-or

firm

-gen

erat

edte

xt

�M

anag

ing

serv

ice

rela

tionsh

ips

�R

elax

esth

ebag

-of-

word

sas

sum

ption

ofm

ost

text-

min

ing

met

hods

�R

elat

esth

ete

xt

toa

par

ticu

lar

foca

len

tity

�A

dva

nce

sin

text-

min

ing

met

hods

will

offer

new

opport

unitie

sin

mar

keting

�A

ccura

cyofcu

rren

tap

pro

aches

islim

ited

�C

om

ple

xre

lationsh

ips

may

be

diff

icult

toex

trac

t�

Itis

advi

sed

todev

elop

dom

ain-s

pec

ific

sentim

ent

tools

asse

ntim

ent

sign

als

can

vary

from

one

dom

ain

toan

oth

er

�N

etze

ret

al.(2

012)

�T

oubia

and

Net

zer

(2017)

�Bogh

rati

and

Ber

ger

(2019)

aR

efer

ence

appea

rsin

the

Web

Appen

dix

.

10

extract known entities such as brands. However, in more casual

text, such as social media, such signals are less useful. Com-

mon dictionaries include LIWC (Pennebaker et al. 2015), EL

2.0 (Rocklage, Rucker, and Nordgren 2018), Diction 5.0, or

General Inquirer for psychological states and traits (for exam-

ple applications, see Berger and Milkman [2012]; Ludwig et al.

[2013]; Netzer, Lemaire, and Herzenstein [2019]).

Sentiment dictionaries such as Hedonometer (Dodds et al.

2011), VADER (Hutto and Gilbert 2014), and LIWC can be used

to extract the sentiment of the text. One of the major limitations of

the lexical approaches for sentiment analysis commonly used in

marketing is that they apply a “bag of words” approach—meaning

that word order does not matter—and rely solely on the cooccur-

rence of a word of interest (e.g., “brand”) with positive or negative

words (e.g., “great,” “bad”) in the same textual unit (e.g., a

review). While dictionary approaches may be an easy way to

measure constructs and comparability across data sets, machine

learning approaches trained by human-coded data (e.g., Borah and

Tellis 2016; Hartmann et al. 2018; Hennig-Thurau, Wiertz, and

Feldhaus 2015) tend to be the most accurate way of measuring

suchconstructs (Hartmann et al. 2019), particularly if theconstruct

is complex or the domain is uncommon. For this reason, research-

ers should carefully weigh the trade-off between empirical fit and

theoretical commensurability, taking care to validate any diction-

aries used in the analysis (discussed in the next section).

A specific type of entity extraction includes linguistic-type

entities such as part-of-speech tagging, which assigns a linguis-

tic tag (e.g., verb, noun, adjective) to each entity. Most text

analysis tools (e.g., the tm package in R, the Natural Language

Toolkit package in Python) have a built-in part-of-speech tag-

ging tool. If no predefined dictionary exists, or the dictionary is

not sufficient for the extraction needed, one could add hand-

crafted rules to help define entities. However, the list of rules

can become long, and the task of identifying and writing the

rules can be tedious. If the entity extraction by dictionaries or

rules is difficult or if the entities are less defined, machine

learning–supervised classification approaches (e.g., condi-

tional random fields [Netzer et al. 2012], hidden Markov

models) or deep learning (Timoshenko and Hauser 2019) can

be used to extract entities. The limitation of this approach is

that often a relatively large hand-coded training data set needs

to be generated.

To allow for a combination of words, entities can be defined

as a set of consecutive words, often referred to as n-grams,

without attempting to extract the relationship between these

entities (e.g., the consecutive words “credit card” can create

the unigram entities “credit” and “card” as well as the bigram

“credit card”). This can be useful if the researcher is interested

in using the text as input for a predictive model.

If the researcher wishes to extract entities while understanding

the context in which the entities were mentioned in the text (thus

avoiding the limitation of the bag-of-words approach), the emer-

ging set of tools of word2vec or word embedding (Mikolov et al.

2013) can be employed. Word2vec maps each word or entity to a

vector of latent dimensions called embedding vector based on the

words with which each focal word appears. This approach allows

the researcher not only to extract words but also to understand the

similarity between words based on the similarities between the

embedding vectors (or the similarities between the sentences in

which each word appears). Thus, unlike the previous approaches

discussed thus far, word2vec preserves the context in which the

word appeared. While word embedding statistically captures the

context in which a word appears, it does not directly linguistically

“understand” the relationships among words.

Topic modeling. Entity extraction has two major limitations: (1)

the dimensionality of the problem (often thousands of unique

entities are extracted) and (2) the interpretation of many enti-

ties. Several topic modeling approaches have been suggested to

overcome these limitations. Similar to how factor analysis

identifies underlying themes among different survey items,

topic modeling can identify the general topics (described as a

combination of words) that are discussed in a body of text. This

text summarization approach increases understanding of docu-

ment content and is particularly useful when the objective is

insight generation and interpretation rather than prediction

(e.g., Berger and Packard 2018; Tirunillai and Tellis 2014).

In addition, monitoring topics, as opposed to words, makes it

easier to assess how discussion changes over time (e.g., Zhong

and Schweidel 2019).

Methodologically, topic modeling mimics the data-

generating process in which the writer chooses the topic she

wants to write about and then chooses the words to express

these topics. Topics are defined as word distributions that com-

monly co-occur and thus have a certain probability of appear-

ing in a topic. A document is then described as a probabilistic

mixture of topics.

The two most commonly used tools for topic modeling are

LDA (Blei, Ng, and Jordan 2003) and Poisson factorization

(PF; Gopalan, Hofman and Blei 2013). The predominant

approach prior to LDA and PF was the support-vector-

machine latent semantic analysis (LSA) approach. While LSA

is simpler and faster to implement than LDA and PF, it

requires larger textual corpora and often achieves lower accu-

racy levels. Other approaches include building an ontology of

topics using a combination of human classification of docu-

ments as seeding for a machine learning classification (e.g.,

Moon and Kamakura 2017). Whereas LDA is often simpler to

apply than PF, PF has the advantage of not assuming that the

topic probabilities must sum to one. That is, some documents

may have more topic presences than others, and a document

can have multiple topics with high likelihood of occurrence.

In addition, PF tends to be more stable with shorter text.

Buschken and Allenby (2016) relax the common bag-of-

words assumption underlying the traditional LDA model and

leverage the within-sentence dependencies of online reviews.

LDA2vec is another approach to assess topics while account-

ing for the sequence context in which the word appears

(Moody 2016). In the context of search queries, Liu and Tou-

bia (2018) further extend the LDA approach to hierarchical

LDA for cases in which related documents (queries and search

results) are used to extract the topics. Furthermore, the

Berger et al. 11

researcher can use an unsupervised or seeded LDA approach

to incorporate prior knowledge in the construction and inter-

pretation of the topics (e.g., Puranam, Narayan, and Kadiyali

2017; Toubia et al. 2019).

While topic modeling methods often produce very sensible

topics, because topics are selected solely based on a statistical

approach, the selection of the number of topics and the inter-

pretation of some topics can be challenging. It is recommended

to combine statistical approaches (e.g., the perplexity measure,

which is a model fit–based measure) and researcher judgment

when selecting the number of topics.

Relation extraction. At the most basic level, relationships

between entities can be captured by the mere co-occurrence

of entities (e.g., Boghrati and Berger 2019; Netzer et al.

2012; Toubia and Netzer 2017). However, marketing research-

ers are often more interested in identifying textual relationships

among extracted entities, such as the relationships between

products, attributes, and sentiments. Such relationships are

often more relevant for the firm than merely measuring the

volume of brand mentions or even the overall brand sentiment.

For example, researchers may want to identify whether consu-

mers mentioned a particular problem with a specific product

feature. Feldman et al. (2015) and Netzer et al. (2012) provide

such examples by identifying the textual relationships between

drugs and adverse drug reactions that imply that a certain drug

may cause a particular adverse reaction.

Relation extraction also offers a more advanced route to

capture sentiment by providing the link between an entity of

interest (e.g., a brand) and the sentiment expressed, beyond

their mere cooccurrence. Relation extraction based on the

bag-of-words approach, which treats the sentence as a bag

of unsorted words and searches for word cooccurrence, is

limited because the cooccurrence of words may not imply a

relationship. For example, the cooccurrence of a drug (e.g.,

Advil) with a symptom (e.g., headache) may refer to the

symptom as a side effect of the drug or as the effect the drug

is aiming to alleviate. Addressing such relationships requires

identifying the sequence of words and the linguistic relation-

ship among them. There have been only limited applications

of such relation extraction in marketing, primarily due to the

computational and linguistic complexities involved in accu-

rately making such relational inferences from unstructured

data (see, e.g., the diabetes drugs application in Netzer et al.

[2012]). However, as the methodologies used to extract entity

relations evolve, we expect this to be a promising direction for

marketers to take.

The most commonly used approaches for relation extraction

are handwritten relationship rules, supervised machine learning

approaches, and a combination of these approaches. At the

most basic level, the researcher could write a set of rules that

describe the required relationship. An example of such a rule

may be the co-occurrence of product (e.g., “Ford”), attribute

(e.g., “oil consumption”), and problem (e.g., “excessive”).

However, such approaches tend to require many handwritten

rules and have low recall (they miss many relations) and thus

are becoming less popular.

A more common approach is to train a supervised machine

learning tool. This could be linguistic agnostic approaches

(e.g., deep learning) or natural language processing (NLP)

approaches that aim to understand the linguistic relationship

in the sentence. Such an approach requires a relatively large

training data set provided by human coders in which various

relationships (e.g., sentiment) are observed. One readily

available tool for NLP-based relationship extraction is the

Stanford Sentence and Grammatical Dependency Parser

(http://nlp.stanford.edu:8080/parser/). The tool identifies the

grammatical role of different words in the sentence to identify

their relationship. For example, to assign a sentiment to a

particular attribute, the parser first identifies the presence of

an emotion word and then, in cases where a subject is present,

automatically assesses if there is a grammatical relationship

(e.g., in the sentence “the hotel was very nice,” the adjective

“nice” relates to the subject “hotel”). As with many off-the-

shelf tools, the validity of the tool for a specific relation

extraction needs to be tested.

Finally, beyond the relations between words/entities within

one document, text can also be investigated across documents

(e.g., online reviews, academic articles). For example, a tem-

poral sequence of documents or a portfolio of documents across

a group or community of communicators can be examined for

interdependencies (Ludwig et al. 2013, 2014).

Text Analysis Metrics

Early work in marketing has tended to summarize unstructured

text with structured proxies for this data. For example, in online

reviews, researchers have used volume (e.g., Godes and Mayzlin

2004; Moe and Trusov 2011); valence, often captured by numeric

ratings that supplement the text (e.g., Godes and Silva 2012; Moe

and Schweidel 2012; Ying, Feinberg and Wedel 2006); and var-

iance, often captured using entropy-type measures (e.g., Godes

and Mayzlin 2004). However, these quantifiable metrics often

mask the richness of the text. Several common metrics are often

used to quantify the text itself, as we explain next.

Count measures. Count measures have been used to measure the

frequency of each entity’s occurrence, entities’ co-occurrence,

or entities’ relations. For example, when using dictionaries to

evaluate sentiment or other categories, researchers often use

the proportion of negative and/or positive words in the docu-

ment, or the difference between the two (Berger and Milkman

2012; Borah and Tellis 2016; Pennebaker et al. 2015; Schwei-

del and Moe 2014; Tirunillai and Tellis 2014). The problem

with simple counts is that longer documents are likely to

include more occurrences of every entity. For that reason,

researchers often focus on the proportions of words in the

document that belong to a particular category (e.g., positive

sentiment). The limitation of this simple measure is that some

words are more likely to appear than others. For example, the

12 Journal of Marketing XX(X)

word “laptop” is likely to appear in almost every review in

corpora that is composed of laptop reviews.

Accuracy measures. When evaluating the accuracy of text mea-

sures relative to human-coded or externally validated docu-

ments, measures of recall and precision are often used.

Recall is the proportion of entities in the original text that the

text-mining algorithm was able to successfully identify (it is

defined by the ratio of true positives to the sum of true positives

and false negatives). Precision is the proportion of correctly

identified entities from all entities identified (it is defined by

the ratio of true positives to the sum of true positives and false

positives). On their own, recall and precision measures are

difficult to assess because an improvement in one often comes

at the expense of the other. For example, if one defines that

every entity in the corpora is a brand, recall for brands will be

perfect (you will never miss a brand if it exists in the text), but

precision will be very low (there will be many false positive

identifications of a brand entity).

To create the balance between recall and precision, one can

use the F1 measure—a harmonic mean of the levels of recall

and precision. If the researcher is more concerned with false

positives than false negatives (e.g., it is more important to

identify positives than negatives), recall and precision can be

weighted differently. Alternatively, for unbalanced data with

high proportions of true or false in the populations, a receiver

operating characteristics curve can be used to reflect the rela-

tionship between true positives and false positives, and the area

under the curve is often used as a measure of accuracy.

Similarity measures. In some cases, the researcher is interested in

measuring the similarity between documents (e.g., Ludwig

et al. 2013). How similar is the language used in two adver-

tisements? How different is a song from its genre? In such

cases, measures such as linguistic style matching, similarity

in topic use (Berger and Packard 2018), cosine similarity, and

the Jaccard index (e.g., Toubia and Netzer 2017) can be used to

assess the similarity between the text of two documents.

Readability measures. In some cases, the researcher is interested

in evaluating the readability of the text. Readability can reflect

the sophistication of the writer and/or the ability of the reader to

comprehend the text (e.g., Ghose and Ipeirotis 2011). Common

readability measures include the Flesch–Kincaid reading ease

and the simple measure of gobbledygook (SMOG) measures.

These measures often use metrics such as average number of

syllables and average number of words per sentence to evaluate

the readability of the text. Readability measures often grade the

text on a 1–12 scale reflecting the U.S. school grade-level

needed to comprehend the text. Common text-mining packages

have built-in readability tools.

The Validity of Text-Based Constructs

While the availability of text has opened up a range of research

questions, for textual data to provide value, one must be able to

establish its validity. Both internal validity (i.e., does text

accurately measure the constructs and the relationship between

them?) and external validity (i.e., do the test-based findings

apply to phenomena outside the study?) can be established in

various ways (Humphreys and Wang 2017). Table 5 describes

how the text analysis can be evaluated to improve different

types of validity (Cook and Campbell 1979).

Internal Validity

Internal validity is often a major threat in the context of text

analysis because the mapping between words and the underly-

ing dimension the research aims to measure (e.g., psychologi-

cal state and traits) is rarely straightforward and can vary across

contexts and textual outlets (e.g., formal news vs. social

media). In addition, given the relatively young field of auto-

mated text analysis, validation of many of the methods and

constructs is still ongoing.

Accordingly, it is important to confirm the internal validity

of the approach used. A range of methods can be adopted to

ensure construct, concurrent, convergent, discriminant, and

causal validity. In general, the approach for ensuring internal

validity is to ensure that the text studied accurately reflects the

theoretical concept or topic being studied, does so in a way that

is congruent with prior literature, is discriminant from other

related constructs, and provides ample and careful evidence for

the claims of the research.

Construct validity. Construct validity (i.e., does the text represent

the theoretical concept?) is perhaps the most important to

address when studying text. Threats to construct validity occur

when the text provides improper or misleading evidence of the

construct. For instance, researchers often rely on existing stan-

dardized dictionaries to extract constructs to ensure that their

work is comparable with other work. However, these diction-

aries may not always fit the particular context. For example,

extracting sentiment from financial reports using sentiment

tools developed for day-to-day language may not be appropri-

ate. Particularly when attempting to extract complex constructs

(e.g., psychological states and traits, relationships between con-

sumers and products, and even sentiment), researchers should

attempt to validate the constructs on the specific application to

ensure that what is being extracted from the text is indeed what

they intended to extract. Construct validity can also be chal-

lenged when homonyms or other words do not accurately

reflect what researchers think they do.

Strategies for addressing threats to construct validity require

that researchers examine how the instances counted in the data

connect to the theoretical concept(s) (Humphreys and Wang

2017). Dictionaries can also be validated using a saturation

approach, pulling a subsample of coded entries and verifying

with a hit rate of approximately 80% (Weber 2005). Another

method is to use input from human coders, as is done to support

machine learning applications (as previously discussed). For

example, one can use Amazon Mechanical Turk workers to

label phrases on a scale from “very negative” to “very positive”

for sentiment analysis and then use these words to create a

Berger et al. 13

weighted dictionary. In many cases, multiple methods for dic-

tionary validation are advisable to ensure that one is achieving

both theoretical and empirical fit. For topic modeling, research-

ers infer topics from a list of cooccurring words. However,

these are theoretical inferences made by researchers. As such,

construct validity is equally important and can be ascertained

using some of the same methods of validation, through satura-

tion and calculating a hit rate through manual analysis of a

subset of the data. When using a classification approach, con-

fusion matrices can be produced to provide details on accuracy,

false positives, and false negatives (Das and Chen 2007).

Concurrent validity. Concurrent validity concerns the way that

the researcher’s operationalization of the construct relates to

prior operationalizations. Threats to concurrent validity often

come when researchers create text-based measures inductively

from the text. For instance, if one develops a topic model from

the text, it will be based on the data set and may not therefore

produce topics that are comparable with previous research. To

address these threats, one should compare the operationaliza-

tion with other research and other data sources. For example,

Schweidel and Moe (2014) propose a measure of brand senti-

ment based on social media text data and validate it by

Table 5. Text Analysis Validation Techniques.

Type of Validity Validation Technique Description of Method for Validation References

Internal ValidityConstruct validity Dictionary validation After draft dictionary is created, pull 10% of the sample

and calculate the hit rate. Measures such as hit rates,precision, and recall can be used to measure accuracy.

Weber (2005)

Have survey participants rate words included in thedictionary. Based on this data, the dictionary can alsobe weighted to reflect the survey data.

Brysbaert, Warriner, andKuperman (2014)a

Have three coders evaluate the dictionary categories. Iftwo of the three coders agree that the word is part ofthe category, include; if not, exclude. Calculate overallagreement.

Humphreys (2010);Pennebaker, Francis, andBooth (2001)a

Saturation Pull 10% of instances coded from the data and calculatethe hit rate. Adjust word list until saturation reaches80% hit rate.

Weber (2005)

Concurrent validity Multiple dictionaries Calculate and compare multiple textual measures of thesame construct (e.g., multiple sentiment measures)

Hartmann et al. (2018)

Comparison of topics Compare with other topic models of similar data sets inother research (e.g., hotel reviews)

Mankad et al. (2016)a

Convergent validity Triangulation Look within text data for converging patterns (e.g.,positive/e emotion correlates with known-positiveattributes); apply Principle Components Analysis toshow convergent groupings of words

Humphreys (2010); Kern et al.(2016)

Multiple operationalizations Operationalize constructs with textual and nontextualdata (e.g., sentiment, star rating)

Ghose et al. (2012)a; Mudambi,Schuff, and Zhang (2014)a

Causal validity Control variables Include variables in the model that address rivalhypotheses to control for these effects

Ludwig et al. (2013)

Laboratory study Replicate focal relationship between the independentvariable and dependent variable in a laboratory setting

Spiller and Belogolova (2016)a;Van Laer et al. (2018)

External ValidityGeneralizability Replication with different

data setsCompare the results from the text analysis with the

results obtained other (possibly non-text-related) datasets

Netzer et al. (2012)

Predict key performancemeasure

Include results from text analysis in regression or othermodel to predict a key outcome (e.g., sales,engagement)

Fossen and Schweidel (2019)

Predictive validity Holdout sample Train model on approximately 80%–90% of the data andvalidate the model with the remaining data. Validationcan be done using k-fold validation, which trains themode on k-1 subsets of the data and predicts for theremaining subset of testing.

Jurafsky et al. (2014)

Robustness Different statisticalmeasures, unitizations

Use different, but comparable, statistical measures oralgorithms (e.g., lift, cosine similarity, Jaccardsimilarity), aggregate at different levels (e.g., day,month)

Netzer et al. (2012)

aReference appears in the Web Appendix.

14 Journal of Marketing XX(X)

comparing it with brand measures obtained through a tradi-

tional marketing research survey. Similarly, Netzer et al.

(2012) compare the market structure maps derived from textual

information with those derived from product switching and

surveys, and Tirunillai and Tellis (2014) compare the topics

they identify with those found in Consumer Reports. When

studying linguistic style (Pennebaker and King 1999), for

example, it is beneficial to use robust measures from prior

literature where factor analysis and other methods have already

been employed to create the construct.

Convergent validity. Convergent validity ensures that multiple

measurements of the construct (i.e., words) all converge to the

same concept. Convergent validity can be threatened when

the measures of the construct do not align or have different

effects. Convergent validity can be enhanced by using several

substantively different measures (e.g., dictionaries) of the

same construct to look for converging patterns. For example,

when studying posts about the stock market, Das and Chen

(2007) compare five classifiers for measuring sentiment,

comparing them in a confusion matrix to examine false posi-

tives. Convergent evidence can also come from creating a

correlation or similarity matrix of words or concepts and

checking for patterns that have face validity. For instance,

Humphreys (2010) looks for patterns between the concept

of crime and negative sentiment to provide convergent evi-

dence that crime is negatively valenced in the data.

Discriminant validity. Discriminant validity, the degree to which

the construct measures are sufficiently different from measures

of other constructs, can be threatened when the measurement of

the construct is very similar to that of another construct. For

instance, measurements of sentiment and emotion in many

cases may not seem different because they are measured using

similar word lists or, when using classification, return the same

group of words as predictors. Strategies for ensuring discrimi-

nant validity entail looking for discriminant rather than con-

vergent patterns and boundary conditions (i.e., when and how

is sentiment different from emotion?). Furthermore, theoretical

refinements can be helpful in drawing finer distinctions. For

example, anxiety, anger, and sadness are different kinds of

emotion (and can be measured via psychometrically different

scales), whereas sentiment is usually measured as positive,

negative, or neutral (Pennebaker et al. 2015).

Causal validity. Causal validity is the degree to which the con-

struct, as operationalized in the data set, is actually the cause of

another construct or outcome, and it is best ascertained through

random assignment in controlled lab conditions. Any number

of external factors can threaten causal validity. However, steps

can be taken to enhance causal validity in naturally occurring

textual data. In particular, rival hypotheses and other explana-

tory factors for the proposed causal relationship can be statis-

tically controlled for in the model. For example, Ludwig et al.

(2013) include price discount in the model when studying the

relationship between product reviews and conversion rate to

control for this factor.

External Validity

To achieve external validity, researchers should attempt to

ensure that the effects found in the text apply outside of the

research framework. Because text analysis often uses natu-

rally occurring data that is often of large magnitude, it tends

have a relatively high degree of external validity relative to,

for example, lab experiments. However, establishing external

validity is still necessary due to threats to validity from sam-

pling bias, overfitting, and single-method bias. For example,

online reviews may be biased due to self-selection among

those who elected to review a product (Schoenmuller, Netzer,

and Stahl 2019).

Predictive validity. Predictive validity is threatened when the

construct, though perhaps properly measured, does not have

the expected effects on a meaningful second variable. For

example, if consumer sentiment falls but customer satisfac-

tion remains high, predictive validity could be called into

question. To ensure predictive validity, text-based constructs

can be linked to key performance measures such as sales (e.g.,

Fossen and Schweidel 2019) or consumer engagement (Ash-

ley and Tuten 2015). If a particular construct has been theo-

retically linked to a performance metric, then any text-based

measure of that construct should also be linked to that perfor-

mance metric. Tirunillai and Tellis (2012) show that the vol-

ume of Twitter activity affects stock price, but they find

mixed results for the predictive validity of sentiment, with

negative sentiment being predictive but positive sentiment

having no effect.

Generalizability can be threatened when researchers base

results on a single data set because it is unknown whether the

findings, model, or algorithm would apply in the same way to

other texts or outside of textual measurements. Generalizability

of the results can be established by viewing the results of text

analysis along with other measures of attitude and behavioral

outcomes. For example, Netzer et al. (2012) test their substan-

tive conclusions and methodology on message boards of both

automobile discussions and drug discussions from WebMD.

Evaluating the external validity and generalizability of the

findings is key, because the analysis of text drawn from a

particular source may not reflect consumers more broadly

(e.g., Schweidel and Moe 2014).

Robustness. Robustness can be limited when there is only one

metric or method used in the model. Researchers can ensure

robustness by using different measures for relationships (e.g.,

Pearson correlation, cosine similarity, lift) and probing results

by relaxing different assumptions. The use of holdout samples

and k-fold cross-validation methods can prevent researchers

from overfitting their models and ensure that relationships

found in the data set will hold with other data as well (Jurafsky

et al. 2014; see also Humphreys and Wang 2017). Probing on

Berger et al. 15

different “cuts” of the data can also help. Berger and Packard

(2018), for example, compare lyrics from different genres, and

Ludwig et al. (2013) include reviews of both fiction and non-

fiction books.

Finally, researchers should bear in mind the limitations of

text itself. There are thoughts and feelings that consumers,

managers, or other stakeholders may not express in text. The

form of communication (e.g., tweets, annual reports) may also

shape the message; some constructs may not be explicit enough

to be measured with automated text analysis. Furthermore,

while textual information can often involve large samples,

these samples may not be representative. Twitter users, for

example, tend to be younger and more educated (Smith and

Anderson 2018). Those who contribute textual information,

particularly in social media, may represent polarized points

of view. When evaluating cultural products or social media,

one should consider the system in which they are generated.

Often viewpoints are themselves filtered through a cultural

system (Hirsch 1986; McCracken 1988) or elevated by an algo-

rithm, and the products make it through this process may share

certain characteristics. For this reason, researchers and firms

should use caution when making attributions on the basis of a

cultural text. It is not necessarily a reflection of reality (Jame-

son 2005) but rather may represent ideals, extremes, or insti-

tutionalized perceptions, depending on the context.

Future Research Agenda

We hope this article encourages more researchers and practi-

tioners to think about how they can incorporate textual data into

their research. Communication and linguistics are at the core of

studying text in marketing. Automated text analysis opens the

black box of interactions, allowing researchers to directly

access what is being said and how it is said in marketplace

communication. The notion of text as indicative of meaning-

making processes creates fascinating and truly novel research

questions and challenges. There are many methods and

approaches available, and there is no space to do all of them

justice. While we have discussed several research streams,

given the novelty of text analysis, there are still ample oppor-

tunities for future research, which we discuss next.

Using Text to Reach Across the Marketing Discipline

Returning to how text analysis can unite the tribes of market-

ing, it is worth highlighting a few areas that have mostly been

examined by one research tradition in marketing where fruitful

cross-pollination between tribes is possible through text anal-

ysis. Brand communities were first identified and studied by

researchers coming from a sociology perspective (Muniz and

O’Guinn 2001). Later, qualitative and quantitative researchers

further refined the concepts, identifying a distinct set of roles

and status in the community (e.g., Mathwick, Wiertz, and De

Ruyter 2007). Automated text analysis allows researchers to

study how consumers in these communities interact at scale

and in a more quantifiable manner—for instance, examining

how people with different degrees of power use language and

predict group outcomes based on quantifiably different

dynamics (e.g., Manchanda, Packard, and Pattabhitamaiah

2015). Researchers can track influence, for example, by inves-

tigating which types of users initiate certain words or phrases

and which others pick up on them. Research could examine

whether people begin to enculturate to the language of the

community over time and predict which individuals may be

more likely to stay or leave on the basis of how well they adapt

to the group’s language (Danescu-Niculescu-Mizil et al. 2013;

Srivastava and Goldberg 2017). Quantitative or machine learn-

ing researchers might capture the most commonly discussed

topics and how these dynamically change over the evolution

of the community. Interpretive researchers might examine how

these terms link conceptually, to find underlying community

norms that lead members to stay. Marketing strategy research-

ers might then use or develop dictionaries to connect these

communities to firm performance and to offer directions for

firms regarding how to keep members participating across dif-

ferent brand communities (or contexts).

The progression can flow the other way as well. Outside of a

few early investigations (e.g., Dichter 1966), word of mouth

was originally studied by quantitative researchers interested in

whether interpersonal communication actually drove individ-

ual and market behavior (e.g., Chevalier and Mayzlin 2006;

Iyengar, Van den Bulte, and Valente 2011). More recently,

however, behavioral researchers have begun to study the under-

lying drivers of word of mouth, looking at why people talk

about and share some stories, news, and information rather than

others (Berger and Milkman 2012; De Angelis et al. 2012; for a

review, see Berger [2014]). Marketing strategy researchers

might track the text of word-of-mouth interactions to predict

the emergence of brand crises or social media firestorms (e.g.,

Zhong and Schweidel 2019) as well as when, if, and how to

respond (Herhausen et al. 2019).

Consumer–firm interaction is also a rich area to examine.

Behavioral researchers could use the data from call centers to

better understand interpersonal communication between con-

sumers and firms and record what drives customer satisfaction

(e.g., Packard and Berger 2019; Packard, Moore, and McFerran

2018). The back-and-forth between customers and agents could

be used to understand conversational dynamics. More quanti-

tative researchers should use the textual features of call centers

to predict outcomes such as churn and even go beyond text to

examine vocal features such as tone, volume, and speed of

speech. Marketing strategy researchers could use calls to

understand how customer-centric a company is or assess the

quality, style, and impact of its sales personnel.

Finally, it is worth noting that different tribes not only have

different skill sets but also often study substantively different

types of textual communication. Consumer-to-consumer com-

munication is often studied by researchers in consumer beha-

vior, whereas marketing strategy researchers more often tend to

study firm-to-consumer and firm-to-firm communication. Col-

laboration among researchers from the different subfields may

allow them to combine these different sources of textual data.

16 Journal of Marketing XX(X)

There is ample opportunity to apply theory developed in one

domain to enhance another. Marketing strategy researchers, for

example, often use transaction economics to study business-to-

business relationships through agency theory, but these

approaches may be equally beneficial when studying

consumer-to-consumer communications.

Broadening the Scope of Text Research

As noted in Table 1, certain text flows have been studied more

than others. A large portion of existing work has focused on

consumers communicating to one another through social

media and online reviews. The relative availability of such

data has made it a rich area of study and an opportunity to

apply text analysis to marketing problems.3 Furthermore, for

this area to grow, researchers need to branch out. This

includes expanding (1) data sources, (2) actors examined, and

(3) research topics.

Expand data sources used. Offline word of mouth, for example,

can be examined to study what people talk about and conversa-

tional dynamics. Doctor–patient interactions can be studied to

understand what drives medical adherence. Text items such as

yearbook entries, notes passed between students, or the text of

speed dating conversations can be used to examine relationship

formation, maintenance, and dissolution. Using offline data

requires carefully transcribing content, which increases the

amount of effort required but opens up a range of interesting

avenues of study. For example, we know very little about the

differences between online recommendations and face-to-face

recommendations, where the latter also include the interplay

between verbal and nonverbal information. Moreover, in the

new era of “perpetual contact” our understanding of cross-

message and cross-channel implications is limited. Research

by Batra and Keller (2016) and Villarroel Ordenes et al.

(2018) suggests that appropriate sequencing of messages mat-

ters; it might similarly matter across channels and modality.

Given the rise of technology-enabled realities (e.g., augmented

reality, virtual reality, mixed reality), assistive robotics, and

smart speakers, understanding the roles and potential differ-

ences between language and nonverbal cues could be achieved

using these novel data sources.

Expand dyads between text producers and text receivers. There are

numerous dyads relevant to marketing in which text plays a

crucial role. We discuss just a few of the areas that deserve

additional research.

Considering consumer–firm interactions, we expect to see

more research leveraging the rich information exchanged

between consumers and firms through call centers and chats

(e.g., Packard and Berger 2019; Packard, Moore, and McFerran

2018). These interactions often reflect inbound communication

between customers and the firm, which can have important

implications for the relationship between parties. In addition,

how might the language used on packaging or in brand mission

statements reflect the nature of organizations and their relation-

ship to their consumers? How might the language that is most

impactful in sales interactions differ from the language that is

most useful in customer service interactions? Research could

also probe how the impact of such language varies across

contexts. The characteristics of language used by consumer

packaged goods brands and pharmaceuticals brands in direct-

to-consumer advertising likely differ. Similarly, the way in

which consumers process the language used in disclosures in

advertisements for pharmaceuticals (e.g., Narayanan, Desiraju,

and Chintagunta 2004) and political candidates (e.g., Wang,

Lewis, and Schweidel 2018) may vary.

Turning to firm-to-firm interactions, most conceptual

frameworks on business-to-business (B2B) exchange relations

emphasize the critical role of communication (e.g., Palmatier,

Dant, and Grewal 2007). Communicational aspects have been

linked to important B2B relational measures such as commit-

ment, trust, dependence, relationship satisfaction, and relation-

ship quality. Yet research on actual, word-level B2B

communication is very limited. For example, very little

research has examined the types of information exchanged

between salespeople and customers in offline settings. The

ability to gather and transcribe data at scale points to important

opportunities to do so. As for within-firm communication,

researchers could study informal communications such as

marketing-related emails, memos, and agendas generated by

firms and consumed by their employees.

Similarly, while a great deal of work in accounting and

finance has begun to use annual reports as a data source (for

a review, see Loughran and McDonald [2016]), marketing

researchers have paid less attention to this area to study com-

munication with investors. Most research has used this data to

predict outcomes such as stock performance and other mea-

sures of firm valuation. Given recent interest in linking

marketing-related activities to firm valuation (e.g., McCarthy

and Fader 2018), this may be an area to pursue further. All firm

communication, including required documents such as annual

reports or discretionary forms of communication such as adver-

tising and sales interactions, can be used to measure variables

such as market orientation, marketing capabilities, marketing

leadership styles, and even a firm’s brand personality.

There are also ample research opportunities in the interac-

tions between consumers, firms, and society. Data about the

broader cultural and normative environment of firms, such as

news media and government reports, may be useful to shed

light on the forces that shape markets. To understand how a

company such as Uber navigates resistance to market change,

for example, one might study transcripts of town hall meetings

and other government documents in which citizen input is

heard and answered. Exogenous shocks in the forms of social

movements such as #metoo and #blacklivesmatter have

affected marketing communication and brand image. One

potential avenue for future research is to take a cultural

3 While readily available data facilitates research, there are downsides to be

recognized, including the representatives of such data and the terms of service

that govern the use of this data.

Berger et al. 17

branding approach (Holt 2016) to study how different publics

define, shape, and advocate for certain meanings in the market-

place. Firms and their brands do not exist in a vacuum, inde-

pendent of the society in which they operate. Yet limited

research in marketing has considered how text can be used to

derive firms’ intentions and actions at the societal level. For

example, scholars have shown how groups of consumers such

as locavores (i.e., people who eat locally grown food; Thomp-

son and Coskuner-Balli 2007), fashionistas (Scaraboto and

Fischer 2012), and bloggers (McQuarrie, Miller, and Phillips

2012) shape markets. Through text analysis, the effect of the

intentions of these social groups on the market can then be

measured and better understood.

Another opportunity for future research is the use of textual

data to study culture and cultural success. Topics such as cul-

tural propagation, artistic change, and the diffusion of innova-

tions have been examined across disciplines with the goal of

understanding why certain products succeed while others fail

(Bass 1969; Boyd and Richerson 1986; Cavalli-Sforza and

Feldman 1981; Rogers 1995; Salganik, Dodds, and Watts

2006; Simonton 1980). While success may be random (Bielby

and Bielby 1994; Hirsch 1972), another possibility is that cul-

tural items succeed or fail on the basis of their fit with con-

sumers (Berger and Heath 2005). By quantifying aspects of

books, movies, or other cultural items quickly and at scale,

researchers can measure whether concrete narratives are more

engaging, whether more emotionally volatile movies are more

successful, whether songs that use certain linguistic features are

more likely to top the Billboard charts, and whether books that

evoke particular emotions sell more copies. While not as

widely available as social media data, more and more data on

cultural items has recently become available. Data sets such as

the Google Books corpus (Akpinar and Berger 2015), song

lyric websites, or movie script databases provide a wealth of

information. Such data could enable analyses of narrative

structure to identify “basic plots” (e.g., Reagan et al. 2016; Van

Laer et al. 2019).

Key Marketing Constructs (That Could Be) Measuredwith Text

Beginning with previously developed ways of representing

marketing constructs can help some researchers address valid-

ity concerns. This section details a few of these constructs to

aid researchers who are beginning to use text analysis in their

work (see the Web Appendix). Using prior operationalization

of a construct can ensure concurrent validity—helping build

the literature in a particular domain—but researchers should

take steps to ensure that the prior operationalization has con-

struct validity with their data set.

At the individual level, sentiment and satisfaction are per-

haps some of the most common measurements (e.g., Buschken

and Allenby, 2016; Homburg, Ehm, and Artz 2015; Herhausen

et al. 2019; Ma, Baohung, and Kekre 2015; Schweidel and Moe

2014) and have been validated in numerous contexts. Other

aspects that may be extracted from text include the authenticity

and emotionality of language, which have also been explored

through robust surveys and scales or by combining multiple

existing measurements (e.g., Mogilner, Kamvar, and Aaker

2011; Van Laer et al. 2019). There are also psychological con-

structs, such as personality type and construal level (Kern et al.

2016; Snefjella and Kuperman 2015), that are potentially use-

ful for marketing researchers and could also be inferred from

the language used by consumers.

Future work in marketing studying individuals might con-

sider measurements of social identification and engagement.

That is, researchers currently have an idea of positive or neg-

ative consumer sentiment, but they are only beginning to

explore emphasis (e.g., Rocklage and Fazio 2015), trust, com-

mitment, and other modal properties. To this end, harnessing

linguistic theory of pragmatics and examining phatics over

semantics could be useful (see, e.g., Villarroel et al. 2017).

Once such work is developed, we recommend that researchers

carefully validate approaches proposed to measure such con-

structs along the lines described previously.

At the firm level, constructs have been identified in firm-

produced text such as annual reports and press releases. Mar-

ket orientation, advertising goals, future orientation, deceitful

intentions, firm focus, and innovation orientation have all

been measured and validated using this material (see Web

Appendix Table 1). Work in organizational studies has a his-

tory of using text analysis in this area and might provide some

inspiration and validation in the study of the existence of

managerial frames for sensemaking and the effect of activists

on firm activities.

Future work in marketing at the firm level could further

refine and diversify measurements of strategic orientation

(e.g., innovation orientation, market-driving vs. market-

driven orientations). Difficult-to-measure factors deep in the

organizational culture, structure, or capabilities may be

revealed in the words the firm, its employees, and external

stakeholders use to describe it (see Molner, Prabhu, and Yadav

[2019]). Likewise, the mindsets and management style of mar-

keting leaders may be discerned from the text they use (see

Yadav, Prabhu, and Chandy [2007]). Firm attributes that are

important outcomes of firm action (e.g., brand value) could

also be explored using text (e.g., Herhausen et al. 2019). In

this case, there is an opportunity to use new kinds of data. For

instance, internal, employee-based brand value could be mea-

sured with text on LinkedIn or Glassdoor. Finally, more subtle

attributes of firm language, including conflict, ambiguity, or

openness, might provide some insight into the effects of man-

agerial language on firm success. For this, it may be useful to

examine less formal textual data of interactions such as

employee emails, salesperson calls, or customer service cen-

ter calls.

Less work in marketing has measured constructs on the

social or cultural level, but work in this vein tends to focus

on how firms fit into the cultural fabric of existing meanings

and norms. For instance, institutional logics and legitimacy

have been measured by analyzing media text, as has the rise

18 Journal of Marketing XX(X)

of brand publics that increase discussion of brands within a

culture (Arvidsson and Caliandro 2016).

At the cultural level, marketing research is likely to maintain

a focus on how firms fit into the cultural environment, but it

may also look to how the cultural environment affects consu-

mers. For instance, measurement of cultural uncertainty, risk,

hostility, and change could benefit researchers interested in the

effects of culture on both consumer and firm effects as well as

the effects of culture and society on government and investor

relationships. Measuring openness and diversity through text

are also timely topics to explore and might inspire innovations

in measurement, focusing on, for example, language diversity

rather than the specific content of language. Important cultural

discourses such as language around debt and credit could also

be better understood through text analysis. Measurement of

gender- and race- related language could be useful in exploring

diversity and inclusion in the way firms and consumers react to

text from a diverse set of writers.

Opportunities and Challenges Provided byMethodological Advances

Opportunities. As the development of text analysis tools

advances, we expect to see new and improved use of these

tools in marketing, which can enable scholars to answer ques-

tions we could not previously address or have addressed only in

a limited manner. Here are a few specific method-driven direc-

tions that seem promising.

First, the vast majority of the approaches used for text anal-

ysis in marketing (and elsewhere) rely on bag-of-words

approaches, and thus, the ability to capture true linguistic rela-

tionships among words beyond their cooccurrence was limited.

However, in marketing we are often interested in capturing the

relationship among entities. For example, what problems or

benefits did the customer mention about a particular feature

of a particular product? Such approaches require capturing a

deeper textual relationship among entities than is commonly

used in marketing. We expect to see future development in

these areas as deep learning and NLP-based approaches enable

researchers to better capture semantic relationships.

Second, in marketing we are often interested in the latent

intention or latent states of writers when creating text, such as

their emotions, personality, and motivations. Most of the

research in this area has relied on a limited set of dictionaries

(primarily the LIWC dictionary) developed and validated to

capture such constructs. However, these dictionaries are often

limited in capturing nuanced latent states or latent states that

may manifest differently across contexts. Similar to advances

made in areas such as image recognition, with the availability

of a large number of human-coded training data (often in the

millions) combined with deep learning tools, we hope to see

similar approaches being taken in marketing to capture more

complex behavioral states from text. This would require an

effort to human-code a large and diverse set of textual corpora

for a wide range of behavioral states. Transfer learning meth-

ods commonly used in deep learning tools such as conventional

neural nets can then be used to apply the learning from the

more general training data to any specific application.

Third, there is also the possibility of using text analysis to

personalize customer–firm interactions. Using machine learn-

ing, text analysis can also help personalize the customer inter-

action by detecting consumer traits (e.g., personality) and states

(e.g., urgency, irritation) and perhaps eventually predicting

traits associated with value to the firm (e.g., customer lifetime

value). After analysis, firms can then tailor customer commu-

nication to match linguistic style and perhaps funnel consumers

to the appropriate firm representative. The stakes of making

such predictions may be high, mistakes costly, and there are

clearly contexts in which using artificial intelligence impedes

constructing meaningful customer–firm relationships (e.g.,

health care; Longoni, Bonezzi, and Morewedge 2019).

Fourth, while our discussion has focused on textual con-

tent, text is just one example of unstructured data, with audio,

video, and image being others. Social media posts often marry

text with images or videos. Print advertising usually overlays

text on a carefully constructed visual. Although television

advertising may not include text on the screen, it may have

an audio track that contains text that progresses simultane-

ously with the video.

Until recently, text data has received the most attention,

mainly due to the presence of tools to extract meaningful fea-

tures. That said, tools such as Praat (Boersma 2001) allow

researchers to extract information from audio (e.g., Van Zant

and Berger 2019). One of the advantages of audio data over text

data is that it provides richness in the form of tone and voice

markers that can add to the actual words expressed (e.g., Xiao,

Kim, and Ding 2013). This enables researchers to study not just

what was said, but how it was said, examining how pitch, tone,

and other vocal or paralinguistic features shape behavior.

Similarly, recent research has developed approaches to ana-

lyze images (e.g., Liu, Xuan et al. 2018), either characterizing

the content of the image or identifying features within an

image. Research into the impact of the combination of text and

images is sparse (e.g., Hartmann et al. 2019). For example,

images can be described in terms of their colors. In the context

of print advertising, textual content may be less persuasive

when used in conjunction with images of a particular color

palette, whereas other color palettes may enhance the persua-

siveness of text. Used in conjunction with simple images, the

importance of text may be quite pronounced. But, when text is

paired with complex imagery, viewers may attend primarily to

the image, diminishing the impact of the text. If this is the case,

legal disclosures that are part of an advertisement’s fine print

may not attract the audience’s attention.

Analogous questions arise as to the role that text plays when

incorporated into videos. Research has proposed approaches to

characterize video content (e.g., Liu et al. 2018). In addition to

comprising the script of the video, text may also appear

visually. In addition to the audio context in which text appears,

its impact may depend on the visuals that appear simultane-

ously. It may also be the case that its position within a video,

relative to the start of the video, may moderate its

Berger et al. 19

effectiveness. For example, emotional text content that is spo-

ken later in a video may be less persuasive for several reasons

(e.g., the audience may have ceased paying attention by the

time the text is spoken). Alternatively, the visuals with which

the audio is paired may be more compelling to viewers, or the

previous content of the video may have depleted a viewer’s

attentional resources. As our discussion of both images and

videos suggests, text is but one component of marketing com-

munications. Future research must investigate its interplay with

other characteristics, including not only the content in which it

appears but also when it appears (e.g., Kanuri, Chen, and Srid-

har 2018), and in what media.

Challenges. While there are a range of opportunities, textual data

also brings with it various challenges. First is the interpretation

challenge. In some ways, text analysis seems to provide more

objective ways of measuring behavioral processes. Rather than

asking people how much they focused on themselves versus

others when sharing word of mouth, for example, one can count

the number of first-person (e.g., “I”) and second-person (e.g.,

“you”; Barasch and Berger 2014) pronouns, providing what

seems more like ground truth. But while part of this process

is certainly more objective (e.g., the number of different types

of pronouns), the link between such measures and underlying

processes (i.e., what it says about the word-of-mouth transmit-

ter) still requires some degree of interpretation. Other latent

modes of behavior are even more difficult to count. While

some words (e.g., “love”) are generally positive, for example,

how positive they are may depend heavily on idiosyncratic

individual differences as well as the context.

More generally, there is challenge and opportunity in under-

standing the context in which textual information appears.

While early work in this space, particularly research using

entity extraction, asked questions such as how much emotion

is in a passage of text, more accurate answers to that question

take must take context into account. A restaurant review may

contain lots of negative words, for example, but does that mean

the person hates the food, the service, or the restaurant more

generally? Songs that contain more second person-pronouns

(e.g., “you”) may be more successful (Packard and Berger

2019), but to understand why, it helps to know whether the

lyrics use “you” as the subject or object of the sentence. Con-

text provides meaning, and the more one understands not just

which words are being used but also how they are being used,

the easier it will be to extract insight. Dictionary-based tools

are particularly susceptible to variation in the context in which

the text appears, as dictionaries are often created in a context-

free environment to match multiple contexts. Whenever possi-

ble, it is advised to use a dictionary that was created for the

specific context of study (e.g., the financial sentiment tool

developed by Loughran and McDonald [2016]).

As mentioned previously, there are also numerous metho-

dological challenges. Particularly when exploring the “why,”

hundreds of features can be extracted, making it important to

think about multiple hypothesis testing (and use of Bonferroni

and other corrections). Only the text used by the text creator is

available, so in some sense there is self-selection. Both the

individuals who decide to contribute and the topics people

decide to raise in their writing may suffer from self-selection.

Particularly when text is used to measure (complex) behavioral

constructs, validity of the constructs needs to be considered. In

addition, for most researchers, analyzing textual information

requires retooling and learning a whole new set of skills.

Data privacy challenges represent a significant concern.

Research often uses online product reviews and sales ranking

data scraped from websites (e.g., Wang, Mai, and Chiang 2013)

or consumers’ social media activity scraped from the platform

(e.g., Godes and Mayzlin 2004; Tirunillai and Tellis 2012).

Although such approaches are common, legal questions have

started to arise. LinkedIn was unsuccessful in its attempt to

block a startup company from scraping data that was posted

on users’ public profiles (Rodriguez 2017). While scraping

public data may be permissible under the law, it may conflict

with the terms of service of those platforms that have data of

interest to researchers. For example, Facebook deleted

accounts of companies that violated its data-scraping policies

(Nicas 2018).4 Such decisions raise important questions about

the extent to which digital platforms can control access to

content that users have chosen to make publicly available.

As interest in extracting insights from digitized text and

other forms of digitized content (e.g., images, videos) grows,

researchers should ensure that they have secured the appropri-

ate permissions to conduct their work. Failure to do so may

result in it becoming more difficult to conduct such projects.

One potential solution is the creation of an academic data set,

such as that made available by Yelp (https://www.yelp.com/

dataset), which may contain outdated or scrubbed data to

ensure that it does not pose any risk to the company’s opera-

tions or user privacy.

The collection and analysis of digitized text, as well as other

user-created content, also raises questions around users’ expecta-

tions for privacy. In the wake of the European Union’s General

Data Protection Regulation and revelations about Cambridge

Analytica’s ability to collect user data from Facebook, research-

ers must be mindful of the potential abuses of their work. We

should also consider the extent to which we are overstepping the

intended use of user-generated content. For example, while a user

may understand that actions taken on Facebook may result in their

being targeted with specific advertisements for brands with which

they have interacted, they may not anticipate the totality of their

Facebook and Instagram activity being used to construct psycho-

graphic profiles that may be used by other brands. Understanding

consumers’ privacy preferences with regard to their online beha-

viors and the text they make available could provide important

guidance for practitioners and researchers alike. Another rich area

for future research is the advancement of the precision with which

marketing can be implemented while minimizing intrusions of

privacy (e.g., Provost et al. 2009).

4 Facebook’s terms of service with regard to automated data collection can be

found at https://www.facebook.com/apps/site_scraping_tos_terms.php.

20 Journal of Marketing XX(X)

Concluding Thoughts

Communication is an important facet of marketing that encom-

passes communication between organizations and their partners,

between businesses and their consumers, and among consumers.

Textual data holds details of these communications, and through

automated textual analysis, researchers are poised to convert this

raw material into valuable insights. Many of the recent advances

in the use of textual data were developed in fields outside of

marketing. As we look toward the future and the role of market-

ers, these recent advancements should serve as exemplars. Mar-

keters are well positioned at the interface between consumers,

firms, and organizations to leverage and advance tools to extract

textual information to address some of the key issues faced by

business and society today, such as the proliferation of misin-

formation, the pervasiveness of technology in our lives, and the

role of marketing in society. Marketing offers an invaluable

perspective that is vital to this conversation, but it will only be

by taking a broader perspective, breaking theoretical and meth-

odological silos, and engaging with other disciplines that our

research can reach its largest possible audience to affect the

public discourse. We hope this framework encourages a reflec-

tion on the boundaries that have come to define marketing and

opens avenues for future groundbreaking insights.

Editors

Christine Moorman and Harald van Heerde

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to

the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, author-

ship, and/or publication of this article.

References

Akpinar, Ezgi, and Jonah Berger (2015), “Drivers of Cultural Success:

The Case of Sensory Metaphors,” Journal of Personality and

Social Psychology, 109 (1), 20–34.

Alessa, Ali, and Miad Faezipour (2018), “A Review of Influenza

Detection and Prediction Through Social Networking Sites,” The-

oretical Biology and Medical Modelling, 15 (1), 1–27.

Anderson, Eric T., and Duncan I. Simester (2014), “Reviews Without

a Purchase: Low Ratings, Loyal Customers, and Deception,” Jour-

nal of Marketing Research, 51 (3), 249–69.

Arsel, Zeynep, and Jonathan Bean (2013), “Taste Regimes and

Market-Mediated Practice,” Journal of Consumer Research, 39

(5), 899–917.

Arvidsson, Adam, and Alessandro Caliandro (2016), “Brand Public,”

Journal of Consumer Research, 42 (5), 727–48.

Ashley, Christy, and Tracy Tuten (2015), “Creative Strategies in

Social Media Marketing: An Exploratory Study of Branded Social

Content and Consumer Engagement,” Psychology & Marketing,

32 (1), 15–27.

Barasch, Alixandra, and Jonah Berger (2014), “Broadcasting and Nar-

rowcasting: How Audience Size Affects What People Share,”

Journal of Marketing Research, 51 (3), 286–99.

Bass, Frank M. (1969), “A New Product Growth for Model Consumer

Durables,” Management Science, 15 (5), 215–27.

Batra, Rajeev, and Kevin L. Keller (2016), “Integrating Marketing

Communications: New Findings, New Lessons, and New Ideas,”

Journal of Marketing, 80 (6), 122–45.

Berger, Jonah (2014), “Word of Mouth and Interpersonal Communi-

cation: A Review and Directions for Future Research,” Journal of

Consumer Psychology, 24 (4), 586–607.

Berger, Jonah, and Chip Heath (2005), “Idea Habitats: How the Pre-

valence of Environmental Cues Influences the Success of Ideas,”

Cognitive Science, 29 (2), 195–221.

Berger, Jonah, Yoon Duk Kim, and Robert Meyer (2019), “Emotional

Volatility and Cultural Success,” working paper.

Berger, Jonah, and Katherine L. Milkman (2012), “What Makes Online

Content Viral?” Journal of Marketing Research, 49 (2), 192–205.

Berger, Jonah, Wendy W. Moe, and David A. Schweidel (2019),

“What Makes Stories More Engaging? Continued Reading in

Online Content,” working paper.

Berger, Jonah, and Grant Packard (2018), “Are Atypical Things More

Popular?” Psychological Science, 29 (7), 1178–84.

Berman, Ron, Shiri Melumad, Colman Humphrey, and Robert Meyer

(2019), “A Tale of Two Twitterspheres: Political Microblogging

During and After the 2016 Primary and Presidential Debates,”

Journal of Marketing Research, 56 (6), doi:10.1177/00222437

19861923.

Bielby, William, and Denise Bielby (1994), “‘All Hits Are Flukes’:

Institutionalized Decision Making and the Rhetoric of Network

Prime-Time Program Development,” American Journal of Sociol-

ogy, 99 (5), 1287–1313.

Blei, David M., Andrew Y. Ng, and Michael I. Jordan (2003), “Latent

Dirichlet Allocation,” Journal of Machine Learning Research, 3,

993–1022.

Boersma, Paul (2001), “Praat, a System for Doing Phonetics by

Computer,” Glot International, 5 (9/10), 341–45.

Boghrati, Reihane, and Jonah Berger (2019) “Quantifying 60 Years of

Misogyny in Music,” working paper.

Bollen, Johan, Huina Mao, and Xiaojun Zeng (2011), “Twitter Mood

Predicts the Stock Market,” Journal of Computational Science, 2

(1), 1–8.

Borah, Abhishek, and Gerard J. Tellis (2016), “Halo (Spillover)

Effects in Social Media: Do Product Recalls of One Brand Hurt

or Help Rival Brands?” Journal of Marketing Research, 53 (2),

143–60.

Boyd, Robert, and Peter Richerson (1986), Culture and Evolutionary

Process. Chicago: University of Chicago Press.

Buschken, Joachim, and Greg M. Allenby (2016), “Sentence-Based Text

Analysis for Customer Reviews,” Marketing Science, 35 (6), 953–75.

Cavalli-Sforza, Luigi Luca, and Marcus W. Feldman (1981), Cultural

Transmission and Evolution: A Quantitative Approach. Princeton,

NJ: Princeton University Press.

Chen, Zoey, and Nicholas H. Lurie (2013), “Temporal Contiguity and

Negativity Bias in the Impact of Online Word of Mouth,” Journal

of Marketing Research, 50 (4), 463–76.

Berger et al. 21

Chevalier, Judith A., and Dina Mayzlin (2006), “The Effect of Word

of Mouth on Sales: Online Book Reviews,” Journal of Marketing

Research, 43 (3), 345–54.

Cohn, Michael A., Matthias R. Mehl, and James W. Pennebaker

(2004), “Linguistic Markers of Psychological Change Surrounding

September 11, 2001,” Psychological Science, 15 (10), 687–93.

Cook, Thomas D., and Donald Thomas Campbell (1979), Experimen-

tal and Quasi-Experimental Designs for Generalized Causal Infer-

ence. Boston: Houghton Mifflin.

Danescu-Niculescu-Mizil, Christian, Robert West, Dan Jurafsky, Jure

Leskovec, and Christopher Potts (2013), “No Country for Old

Members: User Lifecycle and Linguistic Change in Online Com-

munities,” in Proceedings of the 22nd International Conference

on World Wide Web. New York: Association for Computing

Machinery, 307–18.

Das, Sanjiv, and Mike Y. Chen (2007), “Yahoo! for Amazon: Senti-

ment Extraction from Small Talk on the Web,” Management Sci-

ence, 53 (9), 1375–88.

De Angelis, Matteo, Andrea Bonezzi, Alessandro M. Peluso,

Derek Rucker, and Michele Costabile (2012), “On Braggarts

and Gossips: A Self-Enhancement Account of Word-of-Mouth

Generation and Transmission,” Journal of Marketing

Research, 49 (4), 551–63.

Dichter, E. (1966), “How Word-of-Mouth Advertising Works,” Har-

vard Business Review, 44 (6), 147–66.

Dodds, Peter Sheridan, Harris Kameron Decker, Isabel M. Kloumann,

Catherine A. Bliss, and Christopher M. Danforth (2011),

“Temporal Patterns of Happiness and Information in a Global

Social Network: Hedonometrics and Twitter,” PLoS ONE, 6

(12), e26752.

Dowling, Grahame R., and Boris Kabanoff (1996), “Computer-Aided

Content Analysis: What Do 240 Advertising Slogans Have in

Common?” Marketing Letters, 7 (1), 63–75.

Eliashberg, Jehoshua, Sam K. Hui, and Z. John Zhang (2007a), “From

Story Line to Box Office: A New Approach for Green-Lighting

Movie Scripts,” Management Science, 53 (6), 881–93.

Eliashberg, Jehoshua, Sam K. Hui, and Z. John Zhang (2007b),

“Assessing Box Office Performance Using Movie Scripts: A

Kernel-Based Approach,” IEEE Transactions on Knowledge &

Data Engineering, 26 (11), 2639–48.

Feldman, Ronen, Oded Netzer, Aviv Peretz, and Binyamin Rosenfeld

(2015), “Utilizing Text Mining on Online Medical Forums to Pre-

dict Label Change Due to Adverse Drug Reactions,” in Proceed-

ings of the 21th ACM SIGKDD International Conference on

Knowledge Discovery and Data Mining. New York: Association

for Computing Machinery, 1779–88.

Fiss, Peer C., and Paul M. Hirsch (2005), “The Discourse of Globa-

lization: Framing and Sensemaking of an Emerging Concept,”

American Sociological Review, 70 (1), 29–52.

Fossen, Beth L., and David A. Schweidel (2019), “Social TV, Adver-

tising, and Sales: Are Social Shows Good for Advertisers?” Mar-

keting Science, 38 (2), 274–95.

Gandomi, Amir, and Murtaza Haider (2015), “Beyond the Hype: Big

Data Concepts, Methods, and Analytics,” International Journal of

Information Management, 35 (2), 137–44.

Garg, Nikhil, Londa Schiebinger, Dan Jurafsky, and James Zou

(2018), “Word Embeddings Quantify 100 Years of Gender and

Ethnic Stereotypes,” Proceedings of the National Academy of

Sciences, 115 (16), E3635–44.

Gebhardt, Gary F., Francis J. Farrelly, and Jodie Conduit (2019),

“Market Intelligence Dissemination Practices,” Journal of Market-

ing, 83 (3), 72–90.

Ghose, Anindya, and Panagiotis G. Ipeirotis (2011), “Estimating the

Helpfulness and Economic Impact of Product Reviews: Mining

Text and Reviewer Characteristics,” IEEE Transactions on Knowl-

edge and Data Engineering, 23 (10), 1498–1512.

Godes, David, and Dina Mayzlin (2004), “Using Online Conversa-

tions to Study Word-of-Mouth Communication,” Marketing Sci-

ence, 23 (4), 545–60.

Godes, David, and Jose C. Silva (2012), “Sequential and Temporal

Dynamics of Online Opinion,” Marketing Science, 31 (3), 448–73.

Goffman, Erving (1959), “The Moral Career of the Mental Patient,”

Psychiatry, 22 (2), 123–42.

Gopalan, Prem, Jake M. Hofman, and David M. Blei (2013), “Scalable

Recommendation with Poisson Factorization,” (accessed August

19, 2019), https://arxiv.org/abs/1311.1704.

Hancock, Jeffrey T., Lauren E. Curry, Saurabh Goorha, and Michael

Woodworth (2007), “On Lying and Being Lied To: A Linguistic

Analysis of Deception in Computer-Mediated Communication,”

Discourse Processes, 45 (1), 1–23.

Hartmann, Jochen, Mark Heitmann, Christina Schamp, and Oded

Netzer (2019), “The Power of Brand Selfies in Consumer-

Generated Brand Images,” working paper.

Hartmann, Jochen, Juliana Huppertz, Christina Schamp, and Mark

Heitmann (2018), “Comparing Automated Text Classification

Methods,” International Journal of Research in Marketing, 36

(1), 20–38.

Hennig-Thurau, Thorsten, Caroline Wiertz, and Fabian Feldhaus

(2015), “Does Twitter Matter? The Impact of Microblogging Word

of Mouth on Consumers’ Adoption of New Movies,” Journal of the

Academy of Marketing Science, 43 (3), 375–94.

Herhausen, Dennis, Stephan Ludwig, Dhruv Grewal, Jochen Wulf,

and Marcus Schogel (2019), “Detecting, Preventing, and Mitigat-

ing Online Firestorms in Brand Communities,” Journal of Market-

ing, 83 (3), 1–21.

Hill, Vanessa, and Kathleen M. Carley (1999), “An Approach to

Identifying Consensus in a Subfield: The Case of Organizational

Culture,” Poetics, 27 (1), 1–30.

Hirsch, Arnold R. (1986) “The Last ‘Last Hurrah’,” Journal of Urban

History, 13 (1), 99–110.

Hirsch, Peer M. (1972) “Processing Fads and Fashions: An

Organization-Set Analysis of Cultural Industry Systems,” Ameri-

can Journal of Sociology, 77 (4), 639–59.

Holt, Douglas (2016), “Branding in the Age of Social Media,” Har-

vard Business Review, 94 (3), 40–50.

Homburg, Christian, Laura Ehm, and Martin Artz (2015),

“Measuring and Managing Consumer Sentiment in an Online

Community Environment,” Journal of Marketing Research, 52 (5),

629–41.

Huang, Karen, Michael Yeomans, Alison W. Brooks, Julia Minson,

and Francesca Gino (2017), “It Doesn’t Hurt to Ask: Question-

22 Journal of Marketing XX(X)

Asking Increases Liking,” Journal of Personality and Social Psy-

chology, 113 (3), 430–52.

Humphreys, Ashlee (2010), “Semiotic Structure and the Legitimation

of Consumption Practices: The Case of Casino Gambling,” Jour-

nal of Consumer Research, 37 (3), 490–510.

Humphreys, Ashlee, and Kathryn A. LaTour (2013), “Framing the

Game: Assessing the Impact of Cultural Representations on Con-

sumer Perceptions of Legitimacy,” Journal of Consumer Research,

40 (4), 773–95.

Humphreys, Ashlee, and Rebecca Jen-Hui Wang (2017), “Automated

Text Analysis for Consumer Research,” Journal of Consumer

Research, 44 (6), 1274–1306.

Hutto, Clayton J., and Eric Gilbert (2014), “VADER: A Parsimonious

Rule-Based Model for Sentiment Analysis of Social Media Text,”

in Proceedings of the Eighth International Conference on Weblogs

and Social Media. Palo Alto, CA: Association for the Advance-

ment of Artificial Intelligence.

Iyengar, Raghuram, Christopher Van den Bulte, and Thomas Valente

(2011), “Opinion Leadership and Social Contagion in New Product

Diffusion,” Marketing Science, 30 (2), 195–212.

Jameson, Fredric (2005), Archaeologies of the Future: The Desire

Called Utopia and Other Science Fictions. New York: Verso.

Jurafsky, Dan, Victor Chahuneau, Bryan R. Routledge, and Noah A.

Smith (2014), “Narrative Framing of Consumer Sentiment in

Online Restaurant Reviews,” First Monday, 19 (4), https://firstmon

day.org/ojs/index.php/fm/article/view/4944/3863.

Kanuri, Vamsi K., Yixing Chen, and Shrihari (Hari) Sridhar (2018),

“Scheduling Content on Social Media: Theory, Evidence, and

Application,” Journal of Marketing, 82 (6), 89–108.

Kern, Margaret L., Gregory Park, Johannes C. Eichstaedt, H. Andrew

Schwartz, Maarten Sap, and Laura K. Smith (2016), “Gaining

Insights from Social Media Language: Methodologies and

Challenges,” Psychological Methods, 21 (4), 507–25.

Kubler, Raoul V., Anatoli Colicev, and Koen Pauwels (2017) “Social

Media’s Impact on Consumer Mindset: When to Use Which Senti-

ment Extraction Tool,” Marketing Science Institute Working Paper

Series 17-122-09.

Kulkarni, Dipti (2014), “Exploring Jakobson’s ‘Phatic Function’ in

Instant Messaging Interactions,” Discourse & Communication, 8

(2), 117–36.

LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton (2015), “Deep

Learning,” Nature, 521 (7553), 436–44.

Lee, Thomas Y., and Eric T. Bradlow (2011), “Automated Marketing

Research Using Online Customer Reviews,” Journal of Marketing

Research, 48 (5), 881–94.

Li, Feng, and Timon C. Du (2011), “Who Is Talking? An Ontology-

Based Opinion Leader Identification Framework for Word-of-

Mouth Marketing in Online Social Blogs,” Decision Support

Systems, 51 (1), 190–97.

Liu, Jia, and Olivier Toubia (2018), “A Semantic Approach for Esti-

mating Consumer Content Preferences from Online Search Quer-

ies,” Marketing Science, 37 (6), 855–1052.

Liu, Liu, Daria Dzyabura, and Natalie Mizik (2018), “Visual Listening

In: Extracting Brand Image Portrayed on Social Media,” working

paper.

Liu, Xuan, Savannah Wei Shi, Thales Teixeira, and Michel Wedel

(2018), “Video Content Marketing: The Making of Clips,” Journal

of Marketing, 82 (4), 86–101.

Ljung, M. (2000), “Newspaper Genres and Newspaper English,” in

English Media Texts Past and Present: Language and Textual

Structure, Friedrich Ungerer, ed. Philadelphia: John Benjamins

Publishing, 131–50.

Longoni, Chiara, Andrea A. Bonezzi, and Carey K. Morewedge

(2019), “Resistance to Medical Artificial Intelligence,” Journal

of Consumer Research, published online May 3, DOI: https://doi.

org/10.1093/jcr/ucz013.

Loughran, Tim, and Bill McDonald (2016). “Textual Analysis in

Accounting and Finance: A Survey,” Journal of Accounting

Research, 54 (4), 1187–1230.

Ludwig, Stephan, Ko De Ruyter, Mike Friedman, Elisabeth C.

Bruggen, Martin Wetzels, and Gerard Pfann (2013), “More Than

Words: The Influence of Affective Content and Linguistic Style

Matches in Online Reviews on Conversion Rates,” Journal of

Marketing, 77 (1), 87–103.

Ludwig, Stephan, Ko De Ruyter, Dominik Mahr, Elisabeth C.

Bruggen, Martin Wetzels, and Tom De Ruyck (2014), “Take Their

Word for It: The Symbolic Role of Linguistic Style Matches in

User Communities,” MIS Quarterly, 38 (4), 1201–17.

Ludwig, Stephan, Tom Van Laer, Ko De Ruyter, and Mike Friedman

(2016), “Untangling a Web of Lies: Exploring Automated Detec-

tion of Deception in Computer-Mediated Communication,” Jour-

nal of Management Information Systems, 33 (2), 511–41.

Ma, Liye, Sun Baohung, and Sunder Kekre (2015), “The Squeaky

Wheel Gets the Grease—An Empirical Analysis of Customer

Voice and Firm Intervention on Twitter,” Marketing Science, 34

(5), 627–45.

Manchanda, Puneet, Grant Packard, and A. Pattabhitamaiah (2015),

“Social Dollars: The Economic Impact of Consumer Participation

in a Firm-Sponsored Online Community,” Marketing Science, 34

(3), 367–87.

Mathwick, Charla, Caroline Wiertz, and Ko De Ruyter (2007), “Social

Capital Production in a Virtual P3 Community,” Journal of Con-

sumer Research, 34 (6), 832–49.

McCarthy, Daniel, and Peter Fader (2018), “Customer-Based Corpo-

rate Valuation for Publicly Traded Noncontractual Firms,” Journal

of Marketing Research, 55 (5), 617–35.

McCombs, Maxwell E., and Donald L. Shaw (1972), “The Agenda-

Setting Function of Mass Media,” Public Opinion Quarterly, 36

(2), 176–87.

McCracken, Grant (1988), Qualitative Research Methods: The Long

Interview. Newbury Park, CA: SAGE Publications.

McQuarrie, Edward F., Jessica Miller, and Barbara J. Phillips (2012),

“The Megaphone Effect: Taste and Audience in Fashion

Blogging,” Journal of Consumer Research, 40 (1), 136–58.

Melumad, Shiri J., J. Jeffrey Inman, and Michael Tuan Pham (2019),

“Selectively Emotional: How Smartphone Use Changes User-

Generated Content,” Journal of Marketing Research, 56 (2),

259–75.

Mikolov, Tomas, Kai Chen, Greg Corrado, and Jeffrey Dean (2013),

“Efficient Estimation of Word Representations in Vector Space,”

(accessed August 19, 2019), https://arxiv.org/abs/1301.3781.

Berger et al. 23

Moe, Wendy W., and David A. Schweidel (2012), “Online Product

Opinions: Incidence, Evaluation, and Evolution,” Marketing Sci-

ence, 31 (3), 372–86.

Moe, Wendy W., and Michael Trusov (2011), “The Value of Social

Dynamics in Online Product Ratings Forums,” Journal of Market-

ing Research, 48 (3), 444–56.

Mogilner, Cassie, Sepandar D. Kamvar, and Jennifer Aaker (2011),

“The Shifting Meaning of Happiness,” Social Psychological and

Personality Science, 2 (4), 395–402.

Molner, Sven, Jaideep C. Prabhu, and Manjit S. Yadav (2019), “Lost in

the Universe of Markets: Toward a Theory of Market Scoping for

Early-Stage Technologies,” Journal of Marketing, 83 (2), 37–61.

Moody, Christopher E. (2016), “Mixing Dirichlet Topic Models and

Word Embeddings to Make lda2vec,” (accessed August 19, 2019),

https://arxiv.org/abs/1605.02019.

Moon, Sangkil, and Wagner A. Kamakura (2017), “A Picture Is Worth

a Thousand Words: Translating Product Reviews into a Product

Positioning Map,” International Journal of Research in Marketing,

34 (1), 265–85.

Moorman, Christine, Harald J. van Heerde, C. Page Moreau, and

Robert W. Palmatier (2019a), “Challenging the Boundaries of

Marketing,” Journal of Marketing, 83 (5), 1–4.

Moorman, Christine, Harald J. van Heerde, C. Page Moreau, and

Robert W. Palmatier (2019b), “ JM as a Marketplace of Ideas,”

Journal of Marketing, 83 (1), 1–7.

Muniz, Albert, Jr., and Thomas O’Guinn (2001), “Brand Commu-

nity,” Journal of Consumer Research, 27 (4), 412–32.

Narayanan, Sridhar, Ramarao Desiraju, and Pradeep K. Chintagunta

(2004), “Return on Investment Implications for Pharmaceutical

Promotional Expenditures: The Role of Marketing-Mix Inter-

actions,” Journal of Marketing, 68 (4), 90–105.

Netzer, Oded, Ronen Feldman, Jacob Goldenberg, and Moshe Fresko

(2012), “Mine Your Own Business: Market-Structure Surveillance

Through Text Mining,” Marketing Science, 31 (3), 521–43.

Netzer, Oded, Alain Lemaire, and Michal Herzenstein (2019), “When

Words Sweat: Identifying Signals for Loan Default in the Text of

Loan Applications,” Journal of Marketing Research, published

electronically August 15, 2019, doi:10.1177/0022243719852959.

Nicas, Jack (2018), “Facebook Says Russian Firms ‘Scraped’ Data,

Some for Facial Recognition,” The New York Times (October 12),

https://www.nytimes.com/2018/10/12/technology/facebook-rus

sian-scraping-data.html.

Nisbett, Richard E., and Timothy D. Wilson (1977), “Telling More

Than We Can Know: Verbal Reports on Mental Processes,” Psy-

chological Review, 84 (3), 231–59.

Opoku, Robert, Russell Abratt, and Leyland Pitt (2006),

“Communicating Brand Personality: Are the Websites Doing the

Talking for the Top South African Business Schools?” Journal of

Brand Management, 14 (1/2), 20–39.

Ott, Myle, Claire Cardie, and Jeff Hancock (2012), “Estimating the

Prevalence of Deception in Online Review Communities,” in Pro-

ceedings of the 21st International Conference on World Wide Web.

New York: Association for Computing Machinery, 201–10.

Packard, Grant, and Jonah Berger (2017), “How Language Shapes

Word of Mouth’s Impact,” Journal of Marketing Research, 54

(4), 572–88.

Packard, Grant, and Jonah Berger (2019), “How Concrete Language

Shapes Customer Satisfaction,” working paper.

Packard, Grant, Sarah G. Moore, and Brent McFerran (2018), “(I’m)

Happy to Help (You): The Impact of Personal Pronoun Use in

Customer–Firm Interactions,” Journal of Marketing Research, 55

(4), 541–55.

Palmatier, Robert W., Rajiv P. Dant, and Dhruv Grewal (2007), “A

Comparative Longitudinal Analysis of Theoretical Perspectives of

Interorganizational Relationship Performance,” Journal of Mar-

keting, 71 (4), 172–94.

Peladeau, N. (2016), WordStat: Content Analysis Module for SIM-

STAT. Montreal, Canada: Provalis Research.

Pennebaker, James W. (2011), “The Secret Life of Pronouns,” New

Scientist, 211 (2828), 42–45.

Pennebaker, James W., Roger J. Booth, Ryan L. Boyd, and Martha E.

Francis (2015), Linguistic Inquiry and Word Count: LIWC2015.

Austin, TX: Pennebaker Conglomerates.

Pennebaker, James W., and Laura A. King (1999), “Linguistic Styles:

Language Use as an Individual Difference,” Journal of Personality

and Social Psychology, 77 (6), 1296–1312.

Smith, Aaron, and Monica Anderson (2018), “Social Media Use in

2018,” Pew Research Center (March 1), http://www.pewinternet.

org/2018/03/01/social-media-use-in-2018/.

Pollach, Irene (2012), “Taming Textual Data: The Contribution of

Corpus Linguistics to Computer-Aided Text Analysis,” Organiza-

tional Research Methods, 15 (2), 263–87.

Provost, Foster, Brian Dalessandro, Rod Hook, Xiaohan Zhang, and

Alan Murray (2009), “Audience Selection for On-Line Brand

Advertising: Privacy-Friendly Social Network Targeting,” in Pro-

ceedings of the 15th ACM SIGKDD International Conference on

Knowledge Discovery and Data Mining. New York: Association

for Computing Machinery, 707–16.

Puranam, Dinesh, Vishal Narayan, and Vrinda Kadiyali (2017), “The

Effect of Calorie Posting Regulation on Consumer Opinion: A

Flexible Latent Dirichlet Allocation Model with Informative

Priors,” Marketing Science, 36 (5), 726–46.

Ransbotham, Sam, Nicholas Lurie, and Hongju Liu (2019), “Creation

and Consumption of Mobile Word of Mouth: How Are Mobile

Reviews Different?” Marketing Science, published online January

28, DOI: https://doi.org/10.1287/mksc.2018.1115.

Reagan, Andrew J., Lewis Mitchell, Dilan Kiley, Christopher M.

Danforth, and Petter Sheridan Dodds (2016), “The Emotional Arcs

of Stories Are Dominated by Six Basic Shapes,” EPJ Data Science,

5 (1), 1–12.

Rocklage, Matthew D., and Russell H. Fazio (2015), “The Evaluative

Lexicon: Adjective Use as a Means of Assessing and Distinguish-

ing Attitude Valence, Extremity, and Emotionality,” Journal of

Experimental Social Psychology, 56 (January), 214–27.

Rocklage, Matthew D., Derek D. Rucker, and Loran F. Nordgren

(2018), “The Evaluative Lexicon 2.0: The Measurement of Emo-

tionality, Extremity, and Valence in Language,” Behavior

Research Methods, 50 (4), 1327–44.

Rodriguez, Salvador (2017), “U.S. Judge Says LinkedIn Cannot Block

Startup from Public Profile Data,” Reuters (August 14), https://

www.reuters.com/article/us-microsoft-linkedin-ruling/u-s-judge-

24 Journal of Marketing XX(X)

says-linkedin-cannot-block-startup-from-public-profile-data-

idUSKCN1AU2BV.

Rogers, Everett M. (1995). Diffusion of Innovations, 4th ed. New

York: The Free Press.

Rosa, Jose Antonio, Joseph F. Porac, Jelena Runser-Spanjol, and

Michael S. Saxon (1999), “Sociocognitive Dynamics in a Product

Market,” Journal of Marketing, 63 (Special Issue), 64–77.

Rude, Stephanie, Eva-Maria Gortner, and James Pennebaker (2004),

“Language Use of Depressed and Depression-Vulnerable College

Students,” Cognition & Emotion, 18 (8), 1121–33.

Salganik, Matthew J., Peter Dodds, and Duncan Watts (2006),

“Experimental Study of Inequality and Unpredictability in an Arti-

ficial Cultural Market,” Science, 311 (5762), 854–56.

Scaraboto, Daiane, and Eileen Fischer (2012), “Frustrated Fashion-

istas: An Institutional Theory Perspective on Consumer Quests for

Greater Choice in Mainstream Markets,” Journal of Consumer

Research, 39 (6), 1234–57.

Schoenmuller, Verena, Oded Netzer, and Florian Stahl (2019), “The

Extreme Distribution of Online Reviews: Prevalence, Drivers and

Implications,” Columbia Business School Research Paper.

Schweidel, David A., and Wendy W. Moe (2014), “Listening in on

Social Media: A Joint Model of Sentiment and Venue Format

Choice,” Journal of Marketing Research, 51 (4), 387–402.

Simonton, Dean Keith (1980), “Thematic Fame, Melodic Originality,

and Musical Zeitgeist: A Biographical and Transhistorical Content

Analysis,” Journal of Personality and Social Psychology, 38 (6),

972–83.

Snefjella, Bryor, and Victor Kuperman (2015), “Concreteness and

Psychological Distance in Natural Language Use,” Psychological

Science, 26 (9), 1449–60.

Srivastava, Sameer B., and Amir Goldberg (2017), “Language as a Win-

dow into Culture,” California Management Review, 60 (1), 56–69.

Stewart, David W., and David H. Furse (1986), TV Advertising: A

Study of 1000 Commercials. Waltham, MA: Lexington Books.

Tausczik, Yla R., and James.W. Pennebaker (2010), “The Psychological

Meaning of Words: LIWC and Computerized Text Analysis Meth-

ods,” Journal of Language and Social Psychology, 29 (1), 24–54.

Tellis, Gerard J., Deborah J. MacInnis, Seshadri Tirunillai, and

Yanwei Zhang (2019), “What Drives Virality (Sharing) of Online

Digital Content? The Critical Role of Information, Emotion, and

Brand Prominence,” Journal of Marketing, 83 (4), 1–20.

Thompson, Craig J., and Gokcen Coskuner-Balli (2007),

“Countervailing Market Responses to Corporate Co-Optation and

the Ideological Recruitment of Consumption Communities,” Jour-

nal of Consumer Research, 34 (2), 135–52.

Timoshenko, Artem, and John R. Hauser (2019), “Identifying Cus-

tomer Needs from User-Generated Content,” Marketing Science,

38 (1), 1–20.

Tirunillai, Seshadri, and Gerard J. Tellis (2012), “Does Chatter Really

Matter? Dynamics of User-Generated Content and Stock

Performance,” Marketing Science, 31 (2), 198–215.

Tirunillai, Seshadri, and Gerard J. Tellis (2014), “Mining Marketing

Meaning from Online Chatter: Strategic Brand Analysis of Big

Data Using Latent Dirichlet Allocation,” Journal of Marketing

Research, 51 (4), 463–79.

Toubia, Olivier, Garud Iyengar, Renee Bunnell, and Alain Lemaire

(2019), “Extracting Features of Entertainment Products: A Guided

LDA Approach Informed by the Psychology of Media Con-

sumption,” Journal of Marketing Research, 56(1), 18–36.

Toubia, Olivier, and Oded Netzer (2017), “Idea Generation, Creativ-

ity, and Prototypicality,” Marketing Science, 36 (1), 1–20.

Tsai, Jeanne L. (2007), “Ideal Affect: Cultural Causes and Behavioral

Consequences,” Perspectives on Psychological Science, 2 (3),

242–59.

Van Laer, Tom, Jennifer Edson Escalas, Stephan Ludwig, and Ellis A.

Van den Hende (2018), “What Happens in Vegas Stays on Tri-

pAdvisor? Computerized Analysis of Narrativity in Online Con-

sumer Reviews,” Journal of Consumer Research, 46 (2), 267–85.

Van Zant, Alex B., and Jonah Berger (2019), “How the Voice

Persuades,” working paper, Rutgers University.

Villarroel Ordenes, Francisco, Dhruv Grewal, Stephan Ludwig, Ko De

Ruyter, Dominik Mahr, and Martin Wetzels (2018), “Cutting

Through Content Clutter: How Speech and Image Acts Drive Con-

sumer Sharing of Social Media Brand Messages,” Journal of Con-

sumer Research, 45 (5), 988–1012.

Villarroel Ordenes, Francisco, Stephan Ludwig, Ko De Ruyter, Dhruv

Grewal, and Martin Wetzels (2017), “Unveiling What Is Written in the

Stars: Analyzing Explicit, Implicit, and Discourse Patterns of Sentiment

in Social Media,” Journal of Consumer Research, 43 (6), 875–94.

Vosoughi, Soroush, Deb Roy, and Sinan Aral (2018), “The Spread of

True and False News Online,” Science, 359 (6380), 1146–51.

Wang, Xin, Feng Mai, and Roger H.L. Chiang (2013), “Database

Submission—Market Dynamics and User-Generated Content

About Tablet Computers,” Marketing Science, 33 (3), 449–58.

Wang, Yanwen, Michael Lewis, and David A. Schweidel (2018), “A

Border Strategy Analysis of Ad Source and Message Tone in

Senatorial Campaigns,” Marketing Science, 37 (3), 333–55.

Weber, Klaus (2005), “A Toolkit for Analyzing Corporate Cultural

Toolkits,” Poetics, 33 (3/4), 227–52.

Wies, Simone, Arvid Oskar Ivar Hoffmann, Jaako Aspara, and Joost

M.E. Pennings (2019), “Can Advertising Investments Counter the

Negative Impact of Shareholder Complaints on Firm Value?”

Journal of Marketing, 83 (4), 58–80.

Xiao, Li, Hye-Jin Kim, and Min Ding (2013), “An Introduction to

Audio and Visual Research and Applications in Marketing,” in

Review of Marketing Research, Vol. 10, Naresh Malhotra, ed.

Bingley, UK: Emerald Publishing, 213–53.

Xiong, Ying, Moonhee Cho, and Brandon Boatwright (2019),

“Hashtag Activism and Message Frames Among Social Movement

Organizations: Semantic Network Analysis and Thematic Analysis

of Twitter During the #MeToo Movement,” Public Relations

Review, 45 (1), 10–23.

Yadav, Manjit S., Jaideep C. Prabhu, and Rajesh K. Chandy (2007),

“Managing the Future: CEO Attention and Innovation Outcomes,”

Journal of Marketing, 71 (4), 84–101.

Ying, Yuanping, Fred Feinberg, and Michel Wedel (2006),

“Leveraging Missing Ratings to Improve Online Recommendation

Systems,” Journal of Marketing Research, 43 (3), 355–65.

Zhong, Ning, and David A. Schweidel (2019), “Capturing Changes in

Social Media Content: A Multiple Latent Changepoint Topic Mod-

el,” working paper, Emory University.

Berger et al. 25